936. How AI can Boost your English with Podcasts ✨🎧 (feat. Alastair Budge) | Luke's ENGLISH Podcast - Learn British English with Luke Thompson

Race the rudders. Race the sails.

Race the sails. Captain, an unidentified ship is approaching.

Over. Roger. Wait. Is that an Enterprise Sales Solution?

Reach sales professionals, not professional sailors.

With LinkedIn ads, you can target the right people by industry, job title and more.

We'll even give you a $100 credit on your next campaign.

Get started today at LinkedIn .com slash results.

Terms and conditions apply.

This podcast is brought to you by Aura.

Imagine waking up to find your bank account drained, bills for loans you never took out, a warrant for your arrest, all because someone stole your identity.

Hackers aren't waiting.

Why are you? That's why we're thrilled to partner with Aura.

Your personal data is a gold mine for hackers, and Aura helps lock it down.

Aura monitors the dark web, blocks data brokers from selling your information, includes a VPN for private browsing, and a password manager to secure your accounts before criminals break in.

For a limited time, AURA is offering our listeners a 14 -day free trial plus a dark web scan to check if your personal information has been leaked.

All for free at aura .com slash safety.

That's aura .com slash safety to sign up and start protecting yourself and your loved ones.

That's A -U -R -A dot com slash safety.

Terms apply. Check the site for details.

You're listening to Luke's English Podcast. For more information, visit teacherluke .co .uk.

Hello, listeners, welcome back to Luke's English Podcast. Here is another episode about artificial intelligence, AI.

In fact, using AI to help you learn English.

We're going to look at some specific ways that AI can help you to do that, in combination with this podcast. As I always say, listening to plenty of podcast episodes, for example, is a fantastic thing to do for your English.

There are other ways that you can push your English further, other ways that can apply yourself to your English learning.

I think that AI can definitely make this easier.

And in this episode, the idea is to show you some specific ways that you can do that.

Now, AI is incredible, right?

We live in an amazing time, thanks to the potential for learning that AI gives us.

But obviously, artificial intelligence has its negative side as well.

There are controversial points relating to this, as I discussed not long ago in episode 909.

We'll talk about some of those things as well.

There'll be a bit of a discussion about that.

But the main point is learning English with AI.

How can you actually do it?

I'm joined today by Alastair Budge from LeonardoEnglish .com.

Alastair has been working away, looking at specific ways that you can use AI as a learning tool.

And he's come up with a number of very useful prompts and exercises.

And he's going to be kind enough to share them with us today.

So there should be lots of very practical tips and things to learn from this episode, so get yourselves ready.

As usual, there's an episode PDF, which contains a lot of notes and prompts and a transcript of this episode, which should be very useful to help you to get the most out of the episode.

You'll find a link in the description.

So, hello Alastair, how are you doing?

Hello Luke, I am doing wonderfully today.

How are you? Yeah, not bad, thank you.

I'm looking forward to this.

I'm looking forward to seeing what you've come up with.

So what have you been doing?

You've been just sort of playing with AI then?

Yeah, I guess like hundreds of millions of people around the world, and nothing particularly original but have been um yeah trying to see how i personally and how i as a as a language learner can get the maximum benefit from uh ai tools and also as a as a podcaster and someone who you know makes podcasts to try and help people improve their english i have been you know experimenting interesting as much as I can with different ways in which you can use AI tools like Chat GBT in conjunction with podcasts to kind of go one level deeper, really, and help you learn more efficiently with podcasts.

So, yeah, I've been doing this for quite a long time and have kind of created lots of prompts that I found to be quite useful, which people can use in conjunction with any podcast episode to help them improve their, you know, listening comprehension, vocabulary, writing, grammar and so on and even a little bit of speaking.

So it's really been super interesting to play around with and also see the kind of evolution of tools like ChatGBT specifically But really, for, for language learning, because they come on a long way, right?

Yeah, definitely. Yeah.

I mean, I guess it's mostly chat GPT.

Although there are lots of other platforms, but I mean, I, I use chat GPT.

I use others as well, like Microsoft Co pilot and Google Gemini and Google notebook Lm.

There are all these different things I've been playing around with them a lot, too.

But yes, if we just talk about chat GPT, first of all, yeah, I've definitely noticed that it has improved.

It keeps getting better.

It keeps learning doesn't it?

It certainly does. And it's, it has become far more useful from the perspective of, of a language learner as well.

But I still think that lots of people are using it for, you know, just to kind of text generation type tasks as in, write some text for me that I can therefore use and often pass off as my own.

And what I've been really going quite deep into is different ways in which you can, yeah, in which you can use, let's just say, chatgpt rather than chatgpt and other AI tools.

I just refer to chatgpt, um, to mean all AI tools for the, for the time being, rather than just adding a disclaimer every single time.

But, uh, yeah. Ways in which you can use ChachiBT to kind of act as a kind of tutor when listening to a podcast episode, and just get a lot more out of it than than you would do otherwise.

So it's been it's been really, really useful for me.

I've been trying it out with lots of listeners, and the responses have been pretty positive so far.

So I'm very excited to share what I found with people listening to your show.

Okay, great. So can you give us an overview of some of the key ideas or key things you're going to show us?

Sure. Well, maybe before I go right into that, I just wanted to take one step back and talk about podcasts in general because I absolutely love podcasts from the point of view of learning a language.

You always see whenever there are interviews with successful language learners, polyglots, whatever.

They almost always mention podcasts.

And clearly, we both believe that podcasts can be very helpful from the point of view of learning English.

But when you listen to a podcast you're kind of on your own really because you are you know got headphones in you are listening to someone talking.

And perhaps they are talking about vocabulary grammar as well or they just giving you an example of good English good native level English.

And then it's kind of up to you to kind of copy that try to try to learn from that.

And if you are just kind of listening and wandering around or doing the dishes, you know, walking the dog or whatever, then you, you know, you're not really doing anything else with that podcast episode.

And I think you are kind of holding yourself back a little bit from the, from its maximum potential.

So how I like to think about Chat LGBT slash AI tools in terms of learning with a podcast is that they can be a little bit like a sort of tutor who's listening to the podcast with you, and kind of guiding you, helping you pinpoint different interesting words and phrases, helping you test your comprehension, helping you...

and encouraging you to do little mini writing exercises, correcting your grammar, that kind of stuff all based around this podcast episode.

And I think it's the kind of thing that can be so helpful because it turns that slightly passive activity of listening into some kind of short active exercises that can help develop your comprehension, your vocabulary, writing, and grammar and so on.

So that's sort of where I see chatGBT slash AI tools kind of fitting into the, into the picture.

And I really think that, uh, it doesn't have to be as kind of time intensive as lots of people kind of listening to this might think because you think, Oh God, now you're saying you got to like go to your computer and fire chatGBT and do all these things.

It's going to be super complicated.

But, really, the things that I would like to talk about today, they might take you kind of 5 or 10 minutes.

So if you're thinking about listening to a podcast episode that might be half an hour, an hour, two hours long, and then adding on a relatively short activity after that, it doesn't take up so much time out of your day.

And I think it can be very helpful in kind of joining the dots between what you've listened to and what you're actually able to produce.

So yeah, I'm excited to talk about that with with LEP in particular, because one thing I was just reading some of the comments on your, on your YouTube videos before we started this.

And one of the things that kind of the very common thing that people seem to say is like, Well, I love how Luke talks, and Like, I wish I could talk like Luke.

And, you know, you've provided the world with, I don't know how many thousands of hours of examples of how Luke talks.

But clearly not everyone who is listening to your podcast is able to talk just like you.

And so I think ChatGPT can be helpful in terms of really kind of teaching people about different kind of expressions that are used in the podcast, helping them understand the kind of language that's used, um, in a way that would just be very difficult to do if you were on your own.

So, yeah. So it's like an assistant sort of thing.

Um, like a, yeah, an assistant, you can think of it like that.

Um, and you're talking about in terms of the process, we'll get into the details in a bit, but you're, you're talking about essentially like, for example, listening to an episode and then after listening, just quickly using chat GPT to just activate your learning a bit more.

So you'd listen to the episode, do something with chat GPT and then kind of do these exercises that you're going to talk about.

So it's kind of that in that order, because some people listening who perhaps aren't so familiar with how it works, might listen to that and think, oh what, so you mean somehow you can incorporate ChatGPT while I'm listening to the episode?

No, you listen to an episode first and then you essentially give ChatGPT the episode you've listened to and ask it to do certain things.

And Alistair's going to tell us what those, you ask it, the specific questions or prompts that you give it, and then you do the little exercises that it presents to you.

That can be either in written form, texting, like a text chat, or it can be in spoken form.

ChatGPTV these days has quite a functional voice chat, voice conversation mode.

Although the free version, I think that's limited on your phone, for example, it will only do that with you for a certain amount of time.

Obviously, if you use the paid version, you get a unlimited amount of talking time with it.

Before we get into that stuff, I want to focus on that as much as possible.

We're going to talk about things like using it to improve your listening comprehension, to develop your vocabulary, to work on your writing, to work on your grammatical accuracy, to practice speaking.

I think other things, like just helping you to be a better listener to the podcast, being able to search my episode archive more effectively, or even having a conversation with ChatGPT after having heard an episode.

Some of those things.

But I feel like I need to address certain things as well.

There's still a stigma attached to using AI.

Have you noticed this?

Do you know what I mean?

People listening might not know what a stigma is, but there's essentially a negative association.

There is still that slight sense of negativity.

Maybe that's never going to go away, because AI is such a complicated thing, and as we know, there are all sorts of controversial things relating to it.

But so I'll give you an example.

Whenever I'm teaching in a classroom with a group of students and I mention ChatGPT or show them something, for example, I ask ChatGPT, and this is what it said, if we're discussing a language point or discussing something, or whatever.

If I go to ChatGPT and show it on the screen, there's a sense in the room that people have a slightly funny reaction to it.

That's the best thing I can say.

Meaning that maybe people laugh at the idea that let's see what ChatGPT says.

It's kind of funny to use it live in the classroom.

Sometimes there's a slight sense of an intake of breath, oh, like that, as if there's something kind of controversial and negative about it.

Like I think also some teachers maybe don't want to admit that they use ChatGPT, or some of us don't want to admit it, why would we not want to admit it?

It's because there's a slight sense of stigma attached to it.

And that would be things like, maybe it's associated with laziness.

It's a sign of laziness, somehow.

Or it's considered, in some way, to be a sign of cheating.

Using chat GPT to do anything is equivalent to cheating or being lazy.

There's also that association with academic cheating.

we've all heard about the stories of universities attempting to assess their students and the students are all using ChatGPT to write their own essays.

So, that's considered to be a very bad thing.

The effects on education and assessment.

There's copyright issues relating to it.

And also, just the quality concerns with ChatGPT and other software.

What it produces, is it actually good?

Is there actually any value or quality to it?

These are all things that I'd like to discuss later on, but I just wanted to mention those things at the beginning.

Also, there's the issue of using ChatGPT as a language learning tool.

I think that it works best when you do it in the right way.

It can be a very powerful tool.

It can definitely help you learn a language effectively and efficiently, but you have to do it right.

And I think doing it wrong or using chat GPT in the wrong way can actually potentially be harmful for your English.

I think that learners of English should think about this.

And what does this mean, using it wrong?

I think examples of using chat GPT in a way that could be harmful include just asking it to translate something directly for you, using it to do all the work of producing some written English which, as you said earlier, Alastair, which you then pass off as your own, just asking chatGPT to write a reply to a certain comment in a comment section, asking it to reply to a certain email, and then just pasting what it says into what you're writing, your reports or essays or whatever.

Um, you know, that's a kind of like short term, uh, way to use it.

Um, it can be a short term solution, but that might not be what's best for your English.

Long -term, um, and what's wrong with this if I can just say a few more things before we get into the, uh, the stuff you've prepared, Alistair.

Um, you know, you might become so used to using AI.

Uh, to correct you or to generate correct text for you, which you then just copy paste into whatever you're writing.

You may become so used to doing that that you end up kind of doubting your own ability or losing your own ability to do it yourself.

So I think that AI, you should use it to train you to be a better producer of English rather than just using it as a thing that just does all the work for you.

Because, ultimately, this can lead to things like a lack of confidence, a lack of ability, which really hits you when you're in a real world situation and AI isn't available.

It could lead you to sort of lose those productive skills which are so vital for producing good English on your own.

And it can lead to passive learning where you accept corrections from AI, because this is another thing people say, oh, you just asked chat GPT to correct your English, which is great.

But again, if you just accept all those corrections without really asking why or without really learning why the errors happened in the first place, then you don't really maybe internalize English correctly.

You don't really internalize the rules or patterns.

Another thing is that it could potentially lead to a loss of your own authentic voice.

it's very important to find your own voice when you're learning a language, I think.

And AI English can be a little bit generic and characterless and a bit overly formal, and it just might lack a certain special personal touch, which I think is important when making human connections.

And also, finally, there are data concerns.

The fact that you end up sharing a lot of your personal data.

Just simply spending all your time with AI means that you spend less time with humans, which might turn you into some sort of weirdo or something like that.

So I want to have a balanced look.

I want to acknowledge that there are negative sides as well, and also acknowledge that I think there's a right way to use AI.

If you're using in the right way, it can really superpower your English, but you have to do it, you know, correctly.

And I think I think what you're going to tell us about is probably a set of really useful things that can benefit your English.

I hope so. And I certainly agree with everything you said, I just add one thing is that I'm, I'm not here to say that AI should replace any kind of human teacher or anything like that.

And if you are in a situation where there's a human tutor who's listening to episodes of Luke's English podcast with you and providing you with exercises.

That's way better, but that's a much better solution.

For most people that's not a practicality and it's a very expensive thing to hire of thing to hire a private tutor just to set your exercises for Luke's English podcast. Maybe it's a very worthwhile investment, who knows.

I should also add that a lot of the time, my podcast episodes do come with additional material that there's normally some sort of PDF these days, which includes often a vocab list and often different exercises, maybe a vocab quiz that you can use to help you practice, you know, a lot of this time it's all included in the PDF and there's a transcript and stuff like that.

And of course, there's LEP premium as well, which is, you know, stuff that you can use in it's a bit like me teaching you more directly in those episodes.

But yeah, let's have a look at some specific things that people can do.

Alistair, what have you got for us first?

So first, I just wanted to give a little bit of kind of background to how to set this thing, how to set it up really, because what you will need to do before starting working with any episode is you need to give ChatGPT the context of that episode.

And the easiest way of doing that is just by sharing the entire transcripts with ChatGPT.

So maybe I'll just kind of start sharing my screen now.

and I can show people watching this on YouTube how it's working, and for people listening to the audio version of this, I'll just talk you through what I'm doing, so don't worry if you're if you're just listening on audio.

Okay, let me just do this now.

So I've just got chat .gpt up here, I'm going to copy the entire transcript from one of your episodes, episode number 928, which dedicated Luke's English podcast fans will know is The Abduction Learn English with a Short Story.

So I am going to go and I'm going to paste that in there.

Wait, so how do you get the transcript?

Because some people don't know how to do that.

Okay. So I will just - there are automatic subtitles available for episodes.

Either there are non -automatic subtitles, like people have added subtitles into the episodes manually in some way, or YouTube's standard voice -to -text automatic subtitles.

You can access them on pretty much any YouTube video, unless the subtitles have been switched off.

You can access them and can copy and paste them.

And a lot of the time these automatic subtitles are really very accurate, especially these days.

And especially when it's one person talking on their own.

So I'm just showing my screen now, and I'm on the episodes on YouTube.

And I've just gone down here.

And there is a button that says show transcript.

I click on that. And it takes me up to this bit here.

On the right, top right hand corner of the screen, the little window opens up, with the transcript written there.

Exactly. And I'm just going to take the whole thing and I go down here and I'm going to copy it.

So you're just copy pasting the entire thing.

Into ChatGPT. And I'm just going to, I'll just switch my screen, sorry.

And I'm pasting it in there but as I've just pasted it into the chatgpt textbox but an important thing that I will need to do as well is tell chatgpt why I've done this because if I've just given chatgpt a transcript of a podcast episode it won't really know what I'm trying to do it might be a little bit confused and it'll sort of say okay what what do you want me to do with that.

so what I am going to do is I'm going to...

I'm going to paste something else and I'm going to read it out to you and I think Luke this is going to be in the episode PDF so people can just copy it directly from there...

it's this! So above that, within the same sort of chat box as the script, above the transcript you're going to type in your prompt, I've typed in this which is I want to use this podcast transcript to improve my English, I will paste the full transcript below, once I've done that, wait for me to tell you what I want to do next, do not respond until I give you an instruction.

And then below that is just the entire transcript from the podcast episode.

So I will just share this.

I just press send and hopefully okay good.

ChatGPT has come back and said got it let me know what you would like to do with the transcript.

So this is this is a very important step because now ChatGPT has the entire context of the of the podcast episode and we can start to work with it to do whatever we'd like.

Okay, very good. Yeah.

I should also add that, again, you're using a paid version of ChatGPT here.

And if you're using the free version, it might not accept all of that text.

I think it sort of limits it to a certain number of characters.

So you know, you just might you might have to play around with it a little bit and see what the what the limitations are in the free version versus the paid version.

Yeah, you're right, I've got the what's it called, the I think the plus version, which is, I think it's $20 a month, which, for me is is very, very worth it.

I don't want to sound like a chat GPT salesman, but I get a lot of value out of it.

I think with the free plans, you have certain daily limits, I'm not quite sure exactly what they are, and they change all the time.

So perhaps by the time you're watching this, it'll be slightly different.

But the idea is that we want to paste the transcripts of the podcast into chatgpt, and give it the context of why we're doing that.

The episode I chose is a relatively long one.

I think it's an hour and 51 minutes so it's quite long.

But of course you can choose shorter ones.

You can choose to only work with a part of an episode as well.

The important thing is to provide the transcript of what you would like to work with.

Yeah. Okay, so now it's ready.

So you've given it the transcript.

You've said to it, right.

So here's a script, don't do anything yet.

Let me just wait until I tell you what I want to do.

And Chachi PT has said, right, got it.

Let me know what you'd like to do with the script.

And so this is where you add in your next prompt.

So if we use the analogy of, you know, a friendly English tutor.

The tutor has now listened to the episode and it's sort of ready to, um, uh, ready to help us work with it.

So the first category of activities that I'd like to talk about is just to do with listening comprehension because it's the most obvious thing that people might think about developing when they're talking about listening to a podcast. If you listen to a podcast, you develop your listening.

But one of the reasons that I think it's important is that it sort of forces you to listen more actively than you might, otherwise, have listened.

As if you're just wandering along listening while, you know, walking by the river or doing whatever you might be doing, it's easy to not focus on what you're actually listening to.

But if you sort of know that, oh, teacher Chachu PT is going to ask me some questions afterwards, then part of you, especially in me, the kind of lazy learner in me, thinks, OK, well I better, I better really listen carefully if there's going to, there might be some questions coming up.

So that's why I think it's quite a quite a useful thing to do.

Now, I'm going to go back and paste something else in here.

I'll read it out as I as I paste it, but I'm just going to go back to my notes and find the first activity.

so I'm going back into ChatGPT and I have pasted in this and again this is going to be in the PDF I will read it out now so it says you are my English tutor I've just listened to a podcast episode and pasted the transcript above number one ask me five comprehension questions based on the episode to wait for my answer after each one three if I'm wrong or stuck give me a small hint before telling me the correct answer.

And four, after all five, give me a quick summary of how well I understood and suggest one area to improve.

Now before I send this to Chatcheebt, I want just to highlight one thing, which is the third point.

If I'm wrong or stuck, give me a small hint before telling me the correct answer.

Now I think kind of starting to think a bit like this when you're interacting with Chachi BT can be a really useful thing because if you think about what a good teacher would do it's not just immediately tell you the answer or say you're wrong or you're right they would kind of nudge you, kind of push you gently towards getting it for yourself, towards finding out the answer for yourself.

So I think this is a really useful way for people to start thinking about their interactions with ChatGPT, saying, don't just tell me the answer, give me a few hints, help me on my journey to figuring it out for myself and then tell me the answer if I get completely stuck.

So with treating this more as an interaction with a tutor who's helpful, rather than just simple kind of question question and answer uh so I'm gonna go and press uh send and this is obviously live so let's hope it's it's all working uh okay I have to confess that I that I haven't listened to the whole episode so I don't know the answers um see if I can remember let's say so I'll read this out jack peter says great let's get started question one what was the main reason Jose Martinez moved to the UK.

So I think the main reason, so what the learner at this point would do is write in the answer, okay?

So let's say the main reason Jose moved to the UK was to improve his English by immersing himself in a place where English was spoken.

Very good, I will...

To immerse himself in English, let's say.

Okay. Uh, I will just say to immerse, uh, I I'm actually going to add a little grammar, um, mistake there too, to see if chat GPT picks up on it.

I'm going to say, um, to immerse herself in English and see if it corrects me on that after.

Could you add a preposition mistake as well?

So can we say to immerse herself to English?

Why not to, to English.

Okay. okay so what was the main reason Jose Martinez moved to the UK and I am going to answer to immerse herself to English and I'm going to send that now and chatgpt has responded saying almost right you got the idea but there's a small mistake with the pronoun hint remember Jose is a man would you like to try again okay so it's picked up it's given me a hint that I've used the wrong pronoun, that it shouldn't be herself, so I might think okay great it must be himself, but I will keep the same mistake that Luke asked me to add and let's see what happens here.

To immerse himself to English so I've just typed in to immerse himself to English and let's see what happens now.

Okay, close. You've got the right pronoun now, just one small adjustment needed with the preposition.

We don't say immerse to English.

Would you like to try again?

And... Luke, are you okay for me just to put the correct one?

Yeah, go ahead. Okay, we'll go to immerse himself in English.

Okay, and chatgpt has said, perfect!

Yes, Jose moved to the UK to immerse himself in the English language and culture.

Question number two, what unusual event happened to Jose while he was closing up his garage one evening?

Now, I imagine we don't have to go through the entire quiz, but hopefully people listening to this and watching it, can get an idea of how this would work in practice.

We'll go through the entire quiz, and then after question five, ChatGPT will hopefully give us a summary of how well we've done, perhaps highlights any grammar mistakes, corrections with our English, and so on.

So that's just one very simple example of how, if I'd listen to that whole episode.

And I thought, I'll just, I'll just try and test how well I've, how well I've understood it, how I can practice my listening comprehension, and also get a kind of free little grammar, grammar lesson that to based on the mistakes that I've made.

Does it ever feel like you're a marketing professional just speaking into the boy.

Well, with LinkedIn ads, you can know you're reaching the right decision makers.

You can even target buyers by job title, industry, company, seniority, skills.

Wait, did I say job title yet?

Get started today and see how you can avoid the void and reach the right buyers with LinkedIn ads.

We'll even give you a $100 credit on your next campaign.

Get started at LinkedIn .com slash results.

Terms and conditions apply.

This podcast is brought to you by aura.

By the time you hear about a data breach, your information has already been exposed for months.

On average, companies take 277 days to report a breach. That's nine months where hackers have access to your personal data.

That's why we're thrilled to partner with Ora.

Ora is an all -in -one digital safety solution that monitors the dark web for your phone number, email, and social security number, sending real -time alerts if your info is found.

It also includes a VPN, password manager, and data broker removal to help keep you safe.

For a limited time, Ora is offering a 14 -day free trial plus a dark web scan to check if your personal information has been leaked all for free at aura .com slash safety.

That's aura .com slash safety to sign up and protect your loved ones.

That's aura .com slash safety.

Terms apply, check the site for details.

So I would like to say at this point that this shows that as always being a sort of independently minded learner, being a sort of intelligent learner who applies themselves to the learning process, that pays off.

And taking responsibility for your own learning, again, it pays off.

So you might need to take the initiative to, for example, in the case of episode 928, the abduction, you might need to think to yourself, right, well, actually, this is a long episode.

But thinking about it, what Luke has done is he's told us a story, and then he's explain the story, and then he's gone into a lot of vocabulary teaching.

The episode is in those sections.

In fact, often on YouTube there are chapters which show you.

So what you could do is, if you want to focus on listening comprehension, you would take a section of the episode, and you can find the time codes in the transcript.

You can highlight, copy, paste just that section.

For example, here it would be the story of Jose Martinez.

You'd put that into ChatGPT, and then ask it to give you listening comprehension questions based on that.

And that's going to maybe focus you more specifically on your understanding of the story as it was told to you.

And then the vocab teaching part, that's not so much a listening comprehension task for me.

It's more like a question of how much of the vocab did you understand, how many of the things I taught did you remember, and how can you actually practice and use them, you might need to look at other tasks for that section.

But yeah, you need to kind of think in a sort of analytical way a little bit, take the initiative, and, you know, make decisions about how you ask chat GPT to give you listening comprehension practice, namely, choosing just a certain section of the episode to focus on for listening comprehension.

Yeah, precisely. And everything that I will kind of show and talk about today.

I hope listeners will take as sort of...

This is one example of how it can be done.

Now take this and think about it for yourself.

You can just copy and paste the the prompts that will be in the in the episode PDF.

But think really critically about what you're trying to get out of the episode.

If it's one of your short story episodes, then you probably don't need the entire transcript if you're just just listening for kind of comprehension of the story part of the episode.

Yes. Similarly, if you're, if you're working, if you're sort of trying to develop your vocabulary, as we'll talk about, talk about a little bit later on, then you probably want to prompt chatgpt in a way that that tells it that part of the transcript focuses on certain vocabulary words or structures that Luke is talking about, and therefore, that's what you want to focus on.

Yeah, and also, another thing to consider is that you could, if we just take, for example, an IELTS listening exam as an example, one of the things they do there is in the different papers, in the different parts of the exam, they are testing different listening skills.

So for example, different listening skills could be understanding the general meaning of what was said, that's listening for gist, or being able to find specific information from what you listened.

So facts, figures, names, dates, other specific details.

And also understanding the general tone of what you heard. For example, how did the speakers feel about the subject they were talking about?

These are all things that you could ask Chat GPT to test you on.

you know, if it's a conversation with a guest, you could ask it to, um, focus its questions on your ability to understand, let's say the humor that was at play or maybe, um, how the guests felt about the things they were talking about.

Right. Um, so you can ask it to kind of focus on those specific things as well.

Yeah. Okay. Certainly.

And if you know you have the particular area of the, you know, the IELTS listening is a problem area for you then create exercises based just on that particular skill so that you're working on that and not working so much on skills that you're already at a decent level at.

Yeah, that's there.

By the way, you can go to IELTS .org and find preparation resources and a guide for the test, which will tell you, for example, what different skills are being tested by different sections of the exam.

It's quite a good idea to look at that stuff in order to become literate in the language of IELTS, which by extension means becoming literate in the language of you know, the skill set of English you know anyway.

Let's carry on then Alistair.

What else have you got to show us?

I can show you one other kind of quite fun listening comprehension style one.

I can probably add some more prompts that will be in the pdf that we might not talk about in the episode, just so people have got more examples of how you could do this.

But, let's just talk about one more listening comprehension one.

So I'm going to, I'm just going to write into ChatGPT.

Let's stop this particular exercise, and I'm going to paste in something else that I will read out now.

And that is, based on the transcript I provided, choose a short paragraph that contains useful vocabulary and is slightly challenging for a...

I put B2 learner, but of course, people can adapt this as they see fit.

Then I have said, number one, remove six to ten keywords to create a fill in the blank exercise.

Number two, present the blanketed version first and number three, after I try to complete it, show me the correct answers and explain any tricky vocabulary.

So I will just press send there and let's see how we get on.

So you'll see that I'm not going to read out the entire piece of text from Chachi BT, but what it has done is created a fill in the blank exercise.

It's provided us with the timestamp saying approximately 11 minutes and 30 around where Jose encounters the aliens for the first time, and we have the fill in the blank exercise there.

So what this means is it's very easy for me to go back to the podcast episode, just go to 11 minutes and 30, and press play, and then try to complete this fill in the blank exercise.

chat, and of course, like, like before, ChatCPT will act as a tutor, giving us hints.

And it'll explain the harder words for us, and obviously help us with any mistakes that we made.

So what you have to do is, is sort of type out the paragraph, including the missing words, and I'm just looking at the paragraph.

It's like, Jose tried to sit up, but his body felt hmm.

His head was hmm, like he'd just come back from a trip to the pub with his English mate, Barry.

If you remember, listeners, his body felt heavy and his head was spinning.

This is quite good that it's picked out those things, because those are quite meaningful, useful bits of vocab to try to recall.

That's definitely a useful thing to do.

Are there numbers? So, the gaps are not numbered.

But, as you're writing here, Alistair, I can see that you're putting number one, heavy, number two, spinning.

So, you could number them.

ChatGPT is going to understand what that means.

Instead of just typing out the whole paragraph again, you could just put number one, heavy, number two, spinning, and so on.

That's a useful activity.

Those are probably words that I explained as well later on in the episode, I certainly remember explaining his head was spinning.

So that's a useful thing to do.

And it's just randomly picked that paragraph, right?

It just did that randomly.

It just completely randomly picked that paragraph.

I'm sure if listeners tried this at home, they might not get the same, the same paragraph.

I'm sure if you were just trying it now, you probably wouldn't get the same paragraph either.

So it's a kind of useful way to interact in a deeper way than people might otherwise be otherwise be doing.

Yeah. I would say, sorry, if you wanted to, you could pick a certain paragraph and ask it to do that.

You know, input the paragraph and give the same prompt and it'll do that, you know, if there's a specific section that you want to focus on and you could probably ask it to be more specific and say, remove any adjectives or remove all the prepositions.

Or you could say, do you know, if gap out the verbs and I'll have to put them all back in, you know, you could probably get it to, to do those things too, or you just say, just pick out some, as you said, vocab suitable for challenging for a B2 learner or something.

Yeah. Exactly. Should we, should I just, um, should we just see if we can get it to do one of those more specific exercises now to see how it works?

Should we see? OK, so let's just do this live.

Let's move on from this fill -in -the -blank exercise.

Create another fill -in -the -blank exercise that takes a part of the story, not Luke's vocabulary explanations.

Hm. I struggle with, let's say, verbs, should we say?

Yeah, verb tenses. Verbal form, verb forms and would like to focus on this.

I say choose the most complicated part of the story.

How about this? It's find at find a section which includes lots of gerunds and infinitives.

I don't know if it will be able to do that, but let's see.

And then I've said, use the previously defined format for the, um, exercise now listeners, if you're thinking, well, I can't, you have to be very uh, articulate to be able to get chat GPT to do what you want it to do.

Um, but you know, again, check the PDF because a lot of this stuff is written out specifically, and has been already written by Aleister and you can just essentially copy, paste those things.

Yeah, exactly. And hopefully what this demonstration is showing is kind of some of the different ways you can just take these exercises and modify them based on your own goals, your level, and what you want to be doing.

And this is also live, so let's see, let's see what we come up with here.

Okay. I, I just read that out again for the benefit of people listening.

Um, at first I wrote let's move on from this fill in the blank exercise, just to tell just GPT that I wanted to do something else.

Then I said, create another fill in the blank exercise that takes a part of the story brackets not Luke's vocabulary explanations.

I struggle with verb forms and would like to focus on this.

Find a section which includes lots of gerunds and infinitives.

Use the previously defined format for the exercise.

Perhaps, Luke, I should also say and blank out these words.

Yeah. To make sure the chat TPT knows that we want to focus on the gerunds and infinitives and and that's the words we need to find.

It's very specific this one, so I'm curious to see if it will do what I want it to do.

Let's see. Ok, let's just wait for it to finish.

So it says it's done it, here's a fill -in -the -blank exercise focusing on gerunds and infinitives from a story section, not the vocabulary explanations.

It has given us the timestamp approximately 21 minutes 45, when Jose interacts with the aliens and shares jokes.

Can I read this out?

Yes. Yeah. So the paragraph it's chosen is another section from the story.

And it goes like this.

Despite their differences and the fact that they had basically kidnapped him, Jose found himself growing fond of his captors, appreciating their wisdom and the peacefulness of their society.

despite their complete lack of humor and their insistence on _____," and the word is probe.

So, despite their lack of humor and their insistence on _____ as part of their research, he began _____.

The verb is share. He began share stories, music, and literature with them.

So, listeners, what form should these words take?

I'm quite impressed because it has done more or less what I wanted it to do here.

It's despite their insistence on probing him.

That's a gerund form, because normally after a preposition, you would have a gerund form of a verb.

They insisted on probing him.

Despite their insistence on probing him, he began sharing stories.

That's because after the verb begin, we typically, if it's another verb, following it, it would be in the ING form.

They began sharing stories.

But then also, he shared music literature with them along with his recipe for how the perfect paella the verb is make.

How to make the perfect paella.

So yeah, it's done the job.

So if you, listeners, if you do the exercise, you would, you know, just list the verbs in the correct form in the order that they should appear, but you could also ask chat GPT to explain its answers.

You know, you could say, can you give me the answers and explain why they are, how would you put that?

Can you explain the answers?

You could say what I'm doing now, I'm going to write the answers here.

And afterwards, maybe I will also say, please let me know how many I got right and explain any errors that I made.

But what I was also going to do is, Luke, you said that the second one, the answer was he began to share stories.

I said he began sharing but he could be began to share.

Yeah, so I was going to put the wrong one but one that also made grammatical sense and see what happens there.

So, what I've written is, he began to share stories even though in the transcript and the original version it is began sharing and I'm going to say can you Ex, um, oh, so, sorry, how many did I get?

Right. And please explain any, I got wrong.

That's not wonderful English, but let's, um, let's see what it says.

Right. So just to be, it's written quite a lot of stuff here, which is quite a lot of for to read out.

Essentially, the ones that it considered to be grammatically correct, it's given a green tick.

For example, number two, to share.

It says correct. The verb began is followed by the infinitive to share in this context, as it indicates the start of an action.

Now, it doesn't do what I would do as a teacher in this situation when giving feedback on the task, which would be immediately say, yes, you could say began sharing or began to share they're both correct because begin is one of those verbs that actually takes both forms without a difference in meaning but it hasn't actually done that you know maybe it's no big deal but still the one that you got wrong was number five having i don't remember the the specific context of that it was if you go back up to the text it also seems to be quite confused is incorrect the correct answer is having so your

choice is actually correct but I want to explain why just to make sure you're clear.

Right so hold on a minute so it's so you you put for gap number five you put the verb having can you go back up so I can see what that sentence actually was although the zilara showed no signs of yeah it should be having showed no signs of having any sense of humor so having is correct but and you wrote having but it's marked it incorrect so this is like a little bug.

It's kind of made a mistake here.

It said it's incorrect, even though in fact...

so it's instantly contradicted itself.

It says incorrect. The correct answer is HAVING, which is exactly what you wrote.

So your choice is actually correct.

But I want to explain why, just to make sure you're clear.

So Jackgbt's going, you got the right answer, but I just want to make sure that that wasn't pure luck.

The The phrase showed no signs of is followed by a gerund we say show signs of having something or show signs of being something in this context, it means the aliens didn't display any indication of possessing a sense of humor.

So that's odd. It's like it's kind of got marked it wrong, just to kind of make itself feel useful.

So we have a we have a live live live demonstration of how JackGBT, it's not quite there yet.

in terms of getting everything right.

But, it's better than nothing, but always verify that you, you know, because it can make mistakes, as you've seen here.

But what it has done, it hasn't given, the mistake it's made is fairly negligible.

I mean, it's not that serious, because the response it's written is still correct.

It's not teaching you anything that's not correct.

I find that chat GPT and other, uh, you know, generative language models, they almost never make language errors.

I've never, ever seen a grammatical mistake made by a language model like this.

Because it's almost impossible for them to do it because that is the basis of their programming.

Isn't it? It's based on language.

So large language models.

So it's just it's not gonna make mistakes when it comes to language.

This is just some kind of generation error where it thinks it's incorrect but it knows it's actually correct.

So it's got a little bit confused.

But the grammar is right.

Mason, the explanation that is given is correct, but it's just, yeah, it's logic.

This is incorrect. And then it says, yeah, the correct answer.

Actually, your choice is actually correct, but I wanted to just explain it to you anyway because I wanted to feel like I was at least doing something useful i suppose all right there we go okay so we we have those as as a sort of as a few examples of how you can create exercises based on uh listening comprehension do you want to go do you want to do more and talk about kind of vocabulary development stuff or do you want to move into more kind of general discussion of um, of chat GPT.

Let's do a bit of stuff on vocab development, because I think this is a really important and useful thing.

Uh, what we want, right.

Uh, what we want is for chat GPT to help you, to a, uh, help you check that you've understood correctly.

and we've kind of looked at that, but also to help you essentially broaden your range of vocabulary.

And that includes finding exercises, finding ways of helping you to notice, understand, maybe record or remember and recall meaning actually not just remember it in your head, but actually produce new vocabulary, and then actually to personalise it, to use it to express yourself, that's the sort of process that we want.

So let's have a look at some of that.

Sure. What I think I might do is just choose a different episode, given that in this episode, you said there's quite a heavy focus on vocabulary explanations.

So maybe I'll choose one where there are, it's a kind of natural conversation and we can try to, try to use that as a, as a basis for this.

Let's do 921 from comedy to cosy crime with Ian Moore.

Let's try that. 921.

I was talking to a friend of mine who has moved in his career from being a standup comedian to being a writer of crime stories he writes books he's published you find his books in all the major book shops and stuff now so it was a conversation about that okay excellent so what I'm gonna do I'm just going to refresh the chat because chatgbt doesn't need the context of the previous episode and what I'm gonna do is what I before which is I've taken the entire transcript and I'm going to paste in the the kind of preparation prompt which I will just do now so again that is I want to use this podcast

transcript to improve my English I will paste the full transcript below once I've done that wait for me to tell you what I want to do next do not respond until I give you an instruction and And below that, I have just pasted the entire transcript from episode number 921.

So we're going to do that.

I'm what I'm thinking here is that as well as the conversation with Ian in this episode, there was quite a lot of other stuff from me.

So I spent about 15 minutes at the beginning of the episode giving setting a bunch of comprehension questions, like loads of them.

So there's quite a lot of just question asking at the start.

then a conversation.

And then at the end, I spent probably about half an hour in fact, giving answers to all the questions I set at the beginning.

So that's all going to be included as well.

So maybe the smart learner would think, right, well, having listened to that episode, I will choose the script for the conversation.

But whatever, we're going to load in the whole thing and we'll see what happens.

Cool. I'm gonna paste in this prompt here, which I will read out now, which reads based on the transcript I pasted find 10 natural sounding expressions that native speakers use.

But that B1 level learners like me might not.

Of course we can adapt this level as we see fit.

And I've said for each expression, ask me what I think it means, then explain the meaning and how it's typically used, ask me to write a new sentence using the expression and give feedback after each one.

One thing I just wanted to kind of draw listeners attention to here is this, the 10 natural sounding expressions that native speakers use.

but the B1 level learners like me might not because I think this, um, people often tend to think of vocabulary in terms of like, this is a really advanced word and this is a basic word. So, you know, the is a basic word and I don't know, psoriasis or something is a very advanced word that, uh, uh, it's probably also difficult for most people to spell.

Um, uh, But I think there's this kind of almost like second category of words and expressions that are things that are not complicated necessarily, not super advanced or rare, but are the kind of natural expressions that native speakers or advanced English learners will find themselves using a lot.

but you don't often find in kind of textbooks or, um, uh, kind of like formal learning so much and podcasts can be a real sort of gold mine when it comes to these, especially these more conversational style ones.

Um, so in my experience, chatgpt is pretty good at finding these kinds of expressions that people listening to the podcast might not necessarily have picked up on and thought, Oh, that's, that's a really interesting way that Luke used to introduce that subject.

Or, you know, that's, it's the way that his guest kind of explained that in a different way.

That's that that's interesting.

So it's helping people pinpoint these kinds of expressions, explain them, and getting them to use them for themselves.

Yeah, great. Can I just add one, maybe one point in your list, in your prompt?

So you've got, ask me what I think it means.

You know, for each expression, ask me what I think it means, then explain the meaning and how it's typically used.

And third asked me to write a new sentence using the expression.

I feel like I'd like to I'd like it to include the expression in context.

So include a sentence with the expression in context.

That's what want ChatGPT to present to me as a learner.

I don't just want a definition." So it'd be like, I want ChatGPT ultimately to say, here's something that Ian said near the beginning of the conversation, and then it includes a sentence or two containing the phrase.

Because I always think that learners need to see vocab in a grammatical context, or co -text, as it's actually known, they need to see it included in a sentence.

Right, in order to see, because it's vital to see language in its natural environment, if you know what I mean.

That's a very good point.

I think this prompt would provide them the example of the context, the episode.

But what I think could also be useful is we say, make sure you include the word of phrase in the context of the episode, and give another example.

So maybe listeners have the example of how it was used by you or your guests and another example of how it could be used.

So you've got both there.

More examples, the better I think, when learners see a phrase being used again and again and in different sentences, in different contexts, it really helps to kind of provide a meaningful context, which really helps to emphasise the meaning of something.

Okay, I will just read out one more time what I'm asking Chad GPT to do, which is, based on the transcript I pasted, find 10 natural sounding expressions that native speakers use, but the B1 level learners like me might not.

For each expression, ask me what I think it means, then explain the meaning and how it's typically used.

Ask me to write a new sentence using the expression.

And I've written at the bottom make sure that you include the word or phrase in the context of the episode and at least two other examples in different contexts.

Maybe we should have put that after number 2 but I think it should be okay and let's see what ChachiPity does there.

So there's 10 words and phrases.

Okay, it's giving them one by one.

So the first one is fly by the seat of your pants, it's given us the context from the episode, It's asked us to guess what this expression means.

Yeah. Can I read that out?

Yes. Yeah. The first one, it says expression one, fly by the seat of your pants context from the episode.

And this is what something that Ian said, my guest, he said, there tends to be two schools of thinking in terms of crime writers.

There's one school, which are the plotters and the other school which are the pantsers.

It's spelt that wrong.

Because that's pantsers, is actually a kind of made -up word relating to the word pants, p -a -n -t -s, the things that you wear, but it's written in P -A -N -C -E -R -S anyway.

The plotters and the pantsers, they're called the pantsers, spelt correctly now, which means basically, you're just flying by the seat of your pants.

The question is, what do you think, fly by the seat of your pants means?

So just to be clear, two types of two ways to approach writing a crime story.

One is to plot everything.

And the one and the other approach is to fly by the seat of your pants.

You know, it hasn't given our it's given us an explanation.

I think what it's what it's done, which might be, might be quite clever, or perhaps it was the way that I scrolled but it's giving us the expression and then we can if we like I can I can respond to the question before I scroll down and see the explanation yeah um so if I wanted to to really challenge myself I would I'd read that first and I'd say I'm going to really think about this I'm going to write my answer and then I'm going to go and read the explanation which is yeah I mean something's to do.

This expression means to do something without planning using your instinct and improvisation rather than careful preparation.

It comes from aviation where pilots would rely on their instincts rather than instruments to fly the plane, and it's given us two other examples which both seem to be pretty okay to me.

When the surprise guests arrived, I just flew by the seat of my pants and made dinner with whatever I had." It's not something that I would say, but perhaps it's something that you would say, Luke.

Yeah, I would. Yeah.

If just a situation where you haven't prepared and you need to kind of make it up as you go along, you need to improvise, I can imagine.

I think I've explained this, I think I talked about this actually during the episode at some point, maybe in the explanation part at the end, I said something like, you know, imagine going into a classroom without a lesson plan.

And one of your colleagues says, have you not got a lesson plan for that class?

And I'd say, well, there wasn't anything available.

And I don't have time.

Look, I'm just don't worry, I'll just I'm just gonna have to fly by the seat my pants on this one.

I mean, it's quite, it's quite a funny expression.

Yeah, it's quite a cumbersome expression, I might say, don't worry, I'll just improvise, it's probably what would be easier.

I would say that though as well, it depends.

In a very stressed out moment when every second counts as I'm heading into a classroom trying not to be late, I might not say, oh don't worry, I'll just fly by the seat of my pants on this one.

Too many syllables, it's not enough time.

Instead, I might just say, don't worry, I'll improvise.

But anyway, yeah, it's certainly one that I think that people should know.

I think the typo on Panzers, I imagine that's come from the AI -generated transcript rather than chat .gbt writing it wrong.

I think that's probably where it's come from.

Right, that's YouTube's transcript.

You see if you've got the transcript from my website, said that wouldn't have been there.

That mistake wouldn't have been there.

Yeah. Made a rookie mistake there.

Actually, an interesting point to add is that what you can do, as opposed to, uh, going to YouTube and copying down the entire transcripts, you can upload the PDF, I think you provide PDFs Luke.

Um, so you can just upload from computer and you can just say, uh, you know, this is the PDF of the episode, transcript of So episode number 921.

Well, you might need to just be a bit more selective, because often my PDFs don't just have a transcript.

They have a bunch of other stuff, too, including loads of vocab lists and vocab quizzes and things like that as well.

So you might want to open the transcript.

Open up the PDF, find the section which is the transcript.

Often, that's normally the final part.

And just copy -paste all of that, and then paste it into ChatGPT.

see, you know, you do need to think for yourself to a certain extent, listeners.

Yes. He certainly do.

Um, so do you want to see what the next word it comes up with is?

Yeah, let's do one more and then we'll move on to something else.

Should I try and write a sentence and I'm going to try and, um, I'm going to make some grammar mistakes and I'm going to use it in the wrong context.

I fly by the seat of my pants.

So I went on the airplane to France, and I decided to flew by the seat of his pants.

What normally what learners do when they're using idioms is they don't use all of the different parts.

So they'd say I flew by my pants or, uh, or I decided to fly.

By the seat, my pants.

So just forget the, of, okay, let's do this one.

So I went on the airplane to France and I decided to fly by the seat, my pants.

No one knows what that means, but that's all right.

Let's see what chat GPT comes up with.

Let's see what happens here.

Almost there, your sentence is grammatically correct, but the expression, fly by the seat of your pants doesn't literally mean flying on an airplane.

It means doing something without planning or preparation.

Here's a more natural use.

I didn't study much for the exam.

So I had to fly by the seat of my pants during the test. Try rewriting your sentence to show a situation where you're improvising or not well prepared.

Shall I just see what happens if I copy this but I just still make a, make the same grammatical mistake.

Yeah. So with, you're just copying the example, it gave and removing the preposition of that you removed before, but it didn't notice.

And we're going to just see if it is, um, sort of, uh, observant enough, uh, and, um, meticulous enough as a teacher to meet your very particular high standards.

And I have found sometimes the chat GBT will kind of go through one mistake at a time and you'll it'll say you made this one mistake even if you made two mistakes and then you'll fix that mistake but but you'll keep the other mistake and it'll say this is the other mistake which is a little bit frustrating sometimes let's let's see how it how it does in in this example.

So I've written I didn't study much for the exam so I had to fly by the seat my during the test and it has actually done what we wanted.

It's a great job you've got the meaning and context right this time, just one small correction.

Your sentence should be I didn't study much for the exam so I had to fly by the seat of my pants during the test and it's not very clear but of is is bolded here so it's drawing our attention to what we did wrong and it said you just missed the word of after seat other than that is perfect well done.

so it's and then it's obviously remembered the context of the the wider exercise which is the um they're finding native level expressions and the next one is take its toll context from the episode it has really taken its toll physically and mentally and question what do you think take its toll means now hopefully hopefully, as we said before, this is stuff that my listeners will be doing after having heard the episode, right?

So this is after listening stuff, because I said I want it to give the sentence context for the phrase, it has done that, but it's not quite enough.

So all you've got is, it had really taken its toll physically and mentally.

We don't know what it is.

But if you've listened to the episode, you probably would remember the general meaning of what he was talking about.

He was talking about why he...

I think why he quit doing comedy or why he quit touring so much, traveling around, and doing...

I can't remember exactly, but that's a nice expression.

It's a very useful expression.

It's a very common expression.

It's quite a specific expression.

And I think it's exactly the sort of thing that people should learn.

I don't know if it's exactly B1 level.

Normally, at B1 level, I think learners would probably be focusing on slightly more high -frequency language and getting it correct before they move on to this sort of idiomatic language.

But nevertheless, it is definitely the sort of thing that I would find myself teaching in my class, even in an intermediate class.

So to take its toll, it had really taken its toll physically and mentally.

What does take its toll mean?

We could have perhaps changed that prompt to be even more specific.

So rather than saying that language that a B1 learner might not know, we say the kind of language that a B1 learner might not know and should be focusing on for, to achieve that, that B1 level, um, because perhaps I'd be to, um, so perhaps ChatGPT is thinking, okay, this is just something a B1 level learner won't know.

Um, therefore let's just, let's just give it to them rather than focusing on more level appropriate language.

But to be honest, if one of my learners in a B1 class successfully used that phrase properly.

be. For example, if they were saying, I'm gonna give up my job.

Oh, really? Well, to be honest, I'm very tired.

Why? Because traveling so much has really taken its toll.

I think, wow, what's this person doing in an intermediate class?

So to be fair, that is the sort of thing that would get you out of an intermediate class and into a higher level.

So fair enough, I suppose yeah so yeah this is one example of how you can sort of use Chachi BT to help you focus on different words or phrases um and of course in in many of your episodes you focus on these things yourself um so you're you're automatically drawing listeners attention to them and talking about usage and so on.

But in episodes where you're not doing that as much, or if they're more kind of conversational style episodes, this can be a useful way in which you can think about how to use ChatGPT to get some of that value from the audio that you might not necessarily have got otherwise.

Yeah, absolutely. Yeah.

Now I'm sure that you've got loads of other prompts and other things that we could look at, but we can't spend all that much time going through them all.

So we looked at listening comprehension, we looked at focusing on grammar, for example, we've looked at expanding vocab.

I just looked at a few different things.

Is there anything else before we sort of maybe have a discussion about all of this and then end the episode.

So I think there's lots and lots of other things that we could talk about but I think it would probably be a little bit boring for people just to watch me sharing my screen and sort of us battling with chat GPT saying stuff's incorrect when it's actually correct and going through exercises.

What I think is probably the best thing we can do is I can leave a load of those prompts in the PDF and then people can go through them and practice using them, practice adapting them for themselves.

Hopefully what we've done now gives people enough of an idea about how they should be thinking about adapting them for themselves and some of the possibilities of using ChatGBT with podcast episodes.

And of course, how to just set it up from a from a technical point of view.

Mm hmm. Yeah, really good.

That's great. There were actually a couple of other things though that I wanted to ask you about, just a couple of others, before we have our sort of discussion.

And one of those is, you mentioned to me before we started recording that you can use AI to help you search my episode archive because I've got a very long episode archive.

There's like over 900 episodes.

And I'm sure that sometimes my listeners don't quite know where to start if they're looking for episodes where I talk about certain things or whatever.

Maybe AI can help them to find the relevant episode that they're looking for.

Yeah, I mean, I am, I just found it to be, be in many ways a superior way to search the, search, and I kind of archive of anything then Google.

Because if you're saying, if you could go into Google and say, no episodes of Luke's English podcast about travel or something, then it might find some things, but it's probably not going to um, why you might want to, to listen to it.

But if you might, if you said, you know, I, um, I would like to listen to an episode of Luke's English podcast, where Luke talks about his life in France, let's say, um, and it's under an hour long, for example, doesn't exist. You've broken chat, she PT.

Um, but I think, uh, like, um, Like searching for anything.

really, often the results you get with AI just it takes into account better context than, than Google search currently does.

And so I think for the listeners who haven't listened to all 900 and 20, eight, as of the time of recording, Oh, there's by the time this goes up, it'll be 930 watching record. I don't know quite a lot by, if you haven't listened to all of them, Uh, or even if you have, and you want to go back and listen to something about whatever it might be, where Luke talks about cheese or, or no, no, no, no, no, that's good.

I did this whole episode called the cheese episode.

Ah, well, that's true.

Yes. That's probably an example of one.

That's easy to find on Google.

So don't, don't search for that, but, um, you know, once when Luke talks about Luke uses particular types of language that you'd like to, you'd like to focus on.

I think it's just a better way than using Google, because it has all the context, right?

Yeah. It just searches more intelligently than what Google does, which is often it prioritizes certain things.

You know, Google just works in a different way.

But AI might have a more direct approach. You might understand exactly what you're looking for and be able to find it better.

So how do you actually do that, though?

Do you Just go into chat GPT and say, can you recommend episodes of Luke's English podcast where he talks about food?

Exactly. That's all I do.

Um, sorry, it can search the web these days because a couple of years ago when it first came out, it wasn't actually connected to the internet, but these days it does search the web.

Yeah. Yeah. And I think almost every AI model can do, can do that as well.

So whether you're using.

Gemini or grok or deep seafood or anything like that, then they will also be able to do the same thing and provide you with exit with the links.

They can probably also take it one step further and give you some exercises and recommended reading from the BBC or something like that that would compliment it too.

So, um, I would really encourage, yeah, encourage people to just to play around and see what they could find for, um, for themselves.

Um, so let's say, I don't know.

let's say, you are going on a you go on holiday to London, next month, and you are going to be and you're an avid listener of looks English podcast. And you'd like to listen to some to Luke talking about life in London or British customs of London, a tourist hotspots, whatever it might be.

And you'd also like to some relevant articles about London news or things like that and provide a vocabulary list of words that might be useful for your time in London.

I'm not quite sure exactly what you might like but creating a long prompt like that allows you to search all of your archives so you can find an interesting articles, do whatever you want, really.

So, um, yeah, I think it's just something that people should switch, switch to when it comes to looking for particular episodes, I mean, Google is starting to incorporate AI into its search results these days.

So, you know, it won't be long before the old version of Google is replaced by sort of AI powered version, I suppose.

That's just going to become normal.

People will just forget that Google used to be different.

They'll eventually only be familiar with the AI versions of Google, and I think as we move forward, AI is going to become more and more normal and ever -present, to the point where we all kind of won't even really notice that it's It's driving things and powering a lot of the things that we do.

Everything's going to become much easier.

Eventually, our smartphones will have much more powerful AI built into the operating systems, so that it will almost intuitively know what we're trying to get it to do.

Maybe it'll be a thing of the past, where you battle with your phone trying to find certain settings.

Like, how do I reduce the volume of my alarm in this damn thing?

You're trying to get in there.

Eventually you'll just say, Siri reduce the alarm volume, and it'll do it.

It's just going to reduce so much friction in our lives that we'll forget what it used to be like, I expect.

Back to that point, though, of the episode archive and finding episodes, What you could do is just use the URL teacherluke .co .uk slash episodes and give ChatGPT that website address and say search this page for blah, blah blah.

And that's probably going to be a good way of focusing it a bit more.

I think ChatGPT can already do that with the web search. I think that is possible as of as of the time of recording.

Yeah. And the other way, of course, is the old fashioned approach, which is where you open up my episode archive in your browser and then you do a command F or Control F, search for keywords.

And each episode listed on my website has a little summary of what's included.

And that can be a way to, for example, if you're looking for an episode about cheese, or if you just heard me mentioned the cheese episode, you could go into teacherloo .co .uk slash episodes and then control F or command F and then search cheese and the words cheese will be highlighted every time that word has been written on that page.

It'll be highlighted and you can skip to it and eventually you'll find the cheese episode but I mean I can just you know I could just tell you what the number is if you want what is cheese is mentioned once in my episode archive once the cheese episode and that is the cheese episode episode number 72 What is, what is the cheese episode about other than ...

Is it just you talking about cheese?

It's not just, there's a lot of things to say about cheese, especially for someone who lives in France.

There is a lot to say about cheese.

So the, if you open it up that page number 72, the cheese episode, it says, why is this called the cheese episode?

Listen to the whole thing to find out.

And it's actually a rambling episode.

It was a rambling episode with a The rules of the challenge were, I have to speak for 30 minutes.

I'm not allowed to stop speaking or stop the recording unless I get a phone call.

If I get a phone call, I can pause the recording and continue later.

While speaking, I'm not allowed to leave any long pauses.

I have to try to make sense and keep it interesting, which is difficult.

I just talked and talked and talked.

Eventually, I ended up talking about cheese.

That's why I ended up calling it The Cheese Episode.

It's not only about cheese.

It's about a bunch of things, a lot of waffling nonsense, and then some stuff about cheese at the end.

And then I was like, I know, I'll call this the cheese episode.

And everyone will think, why is this called the cheese episode?

And then they'll listen to it, and then they'll discover why.

That was Episode number 70?

72. Recorded. That was published in 2011.

Wow. So that was relatively early in your podcasting journey?

Yes it was. A formative episode where I established the tradition of doing rambling on the podcast. Have your opinions on cheese changed in those 14 years?

My opinions on cheese have matured, you could say.

I have more of an appreciation of cheese I think these days, even more than, you know, obviously, since moving to France, oh, yes, I do like cheese.

I mean, who? Of course, I can't imagine anyone not, vegans of course, vegan cheese.

I've tried it, it wasn't as good as real cheese, I'm afraid.

Although the principle I do agree with is just the taste, not so much. Yeah, and that's the problem.

Yeah. All right. And the other thing I wanted to ask you about was you mentioned before that you could make your own sort of, you can make your own GPT within ChatGPT, and ask it to do certain things.

And am I right in thinking that you've created one that allows people to chat with a kind of AI version of me?

Is this right? Yes, that's correct.

So there are these things called custom GPTs, I think you have to have one of the paid ChatGPT accounts, to be able to create your own, but you can anyone can interact with them, and essentially they are they're kind of like versions of chat gpt that have some pre existing knowledge or instructions, and they can be about anything.

So you can have a bird watching chat, uh, GPT or, uh, GPT about podcasting or, or whatever it might be, or you can have a GPT that is the L -E -P GPT, um, which, which hopefully, uh, which I've tried to create, um, based on your episodes and to act as, as you.

So maybe, and I should also just put a little disclaimer here for anyone thinking, you know, what's going on?

This is this is meant to be more sort of interesting fun and fun, rather than purely educational.

So I thought it'd be fun to make this and say, I wonder what Luke thinks about interacting with an AI version of himself.

and does he think that it's completely ridiculous or has he got something slightly right?

So maybe if you wanted to share your screen and try and you human, Luke, interact with AI, Luke, LEP, GPT, and see what happens.

Okay. So, I should add that this is not like a voice emulation of me, it's not actually my voice because that is a possible thing.

Of course, these days that you can, you know, you can emulate people's voices and all sorts of other incredible things, which are quite frightening.

Like you can create sort of video versions of a person, which is on the more terrifying end of the AI spectrum, where you know, you get deep fakes and all that sort of thing.

There's no actual voice emulation in this GPT that you've created.

It's more like just a text chat that you would have. So I'm going to open it up now.

And I it's also been given instructions not claim that it actually Luke, so there's no sort of impersonation going on here, but hopefully a bit of fun where it might tell some jokes or teach some English or do certain things like that.

Okay, so imagine I'm chatting with myself, what would I ask myself?

Goodness me. This is a bit like when I'm lying around inside my own head.

I'm just going to keep it simple and say, I'm going to start with this, hi, Luke.

All right. And it says, hello, how lovely to hear from you.

How's it going? Got your cuppa ready?

Meaning have you got your cup of tea ready?

Or maybe you're out and about with your headphones on doing that classic English podcast walk thing.

What brings you here today?

Looking for a cheeky language tip?

A podcast episode recommendation?

or maybe just a bit of good old English waffle.

Okay. So this is it's version of me, because I'm always talking about cups of tea and, and things like that.

Um, yeah, I should, I should also also add just, it sounds like I'm trying to give myself a whole lot of disclaimers here, but how it was kind of trained was on your podcast transcripts, rather than me saying, this is how Luke should talk.

Um, so it's not me saying, Luke would definitely say, Got your copper ready.

And what it has done has taken transcripts of the episode and thought, right, this is the persona that I'm going to build.

And then, you know, trying to talk like it thinks this person would.

So that's, that's what is going on.

I got it. So yeah, you fed it lots of my transcripts.

And this is its version of me based on on that.

So got your copper ready is, I suppose, because I'm British, and of course, we drink tea all the time, and a cuppa is a slang word for a cup of tea.

So I can see why it's doing that.

What am I going to ask it?

So have you got any interesting new podcasts in the pipeline?

Oh, now that's a good question.

Always nice to know someone's curious about what's cooking in the LEP kitchen.

Right. So imagine me rubbing my hands together like a mad professor.

Let's see what kind of podcast magic we can conjure up next.

Here are a few things in the pipeline.

Okay. So it's just making stuff up now, cause these are not, it doesn't have some genuine insight somehow by spying on me what actual episodes I have in my actual pipeline, but it's got a deep dive into British humor, blah, blah, blah.

Accents of the UK, part two.

Lepster stories. The lazy grammar guide.

That sounds interesting.

A new grammar focused episode, but delivered Luke style, not like a dusty textbook, but more like here's how native speakers actually use this stuff without thinking about it too much. You'll come away learning loads, but with your brain melting, promise".

Okay, that's not actually...

It's not a bad idea.

The lazy grammar guide, quite like the idea.

You had it here first. Yeah.

Okay. Well, that's a bit of fun.

I suppose you could have fun with that.

The link will be in the PDF for Alister's LEP GPT, which, as he said, is just a bit of fun really.

And it's not entirely accurate.

It's a kind of a, a sort of a version of me, uh, loosely based.

A version that's loosely based in a kind of a caricature version of me.

Let's say, um, that's a, that's quite good fun.

Um, also, this is something I just, just thought about now, but it's probably possible for listeners to create their own GPT that is a kind of, uh, like English learning podcast companion.

So kind of takes in lots of the prompts that we talked about earlier about, you know, how it how they actually behave and so on.

And then you just paste in the transcripts and it is kind of ready to go.

So that's something that people might want to consider doing.

Okay, really good. I said before that we would discuss some stuff before the end of the episode, But I feel like we could and should, um, I mentioned issues earlier on like copyright issues.

Have you, have you seen the recent trend for studio Ghibli imagery in ChatGPT?

Are you aware of the studio Ghibli controversy?

Can you tell us what it is?

So if this, um, animation studio in Japan that has been going for a long time, we're long but you know many decades and as a particular animation style um where i think every single um every single kind of image of the of the animation i think is hand -drawn i believe and i'm i may be getting this slightly wrong but it's an incredibly labour intensive um Process is very distinctive.

So if you see a studio could be, um, uh, image, you kind of immediately recognize it.

Um, and last week, I believe maybe two weeks ago, relatively recently, um, at the end of March or solid April chat GPT.

OpenAI rather released a new image generation model that was just way better than all of the previous ones and allows you to do things like upload an image and say, you know, create this in studio groupie style.

And suddenly everyone was, millions of people around the world were able to create their own images in this very particular style of this animation studio.

And of course, the the animation studio was in no way kind of compensated for it.

And this, I can't remember the name of the man who was kind of the, the boss of the whole thing, and the chief animator.

But he was very, I think, in my opinion, understandably cross about it, because he, you know, spent his life working on creating this style and, um, uh, you know, it's incredibly intensive and expensive to do.

And then suddenly anyone around the world, including me, I should say, um, was able to generate images like that.

And. To me, to my kind of untrained eyes, they looked very, very good, uh, almost unmistakable from, from what it would look like if, uh, if someone asked you Ghibli did that.

So yeah, there's all sorts of copyright, copyright issues.

I also seen, I'm not sure if you've seen this too, Luke.

But people taking some of the kind of transcripts of my episodes and uploading them onto YouTube with different AI generated voices.

So, obviously, I'm in a completely different category to kind of the sizes geographically.

But even making a very small podcast, finding people just take the transcripts and use text generation and then upload it to YouTube to try and pass it off as their own.

Well, people have taken transcripts of your episodes and they fed them into some sort of AI, which then goes from text to speech. Yeah, exactly.

And there, and they, they formed, uh, audio episodes where it's an AI voice, reading out your transcript and they're passing it off as their own content.

Exactly. Yeah. Right.

That's terrible. Yeah.

And that just shows how AI can really make it incredibly easy to sort of violate people's intellectual property in some way to breach people's copyright and also, So yeah, just to create content that they don't have the rights to in some way, it really makes that easier.

I think with something like a podcast, it's in a funny category, I guess, because for someone who is listening to your show, they're here for you, right.

And to much, much lesser extent, the guests who come on your show, but they're here to listen to you and to learn from you.

And so, just kind of slapping on, instead of Luke, it's Ted, whatever, and he's going to talk in a funny accent and not have any of your kind of natural charm and quirks and stuff, that is not going to be popular at all, I would imagine.

But something like an image where, you know, you can't see the process of it being generated where ultimately all you see is the final image in the case of the Studio Ghibli stuff, yeah, it's kind of scary because people can't see how that has been made, and the, the original creator of the concept gets no compensation or attribution for their decades of hard work.

Yeah, yeah. That's the crux of the controversy.

Also, issues of slop.

This is a slang word, a fairly new slang word, which refers to the amount of AI -created content which is everywhere on the internet now.

You talked about someone using transcripts to make their own podcast, using your own transcripts to generate their own podcasts with text -to -speech software, but there are so many, especially on YouTube, so many AI -generated podcasts now, they're very generic.

You've noticed these too, right?

Yeah, and they're huge, some of them.

Some of them are. I was surprised actually because all the ones that I've ever seen, and they pop up in my YouTube recommended videos and stuff, they normally have minimal views, like only a few views really, maybe a few hundred.

But you shared a channel with me which has got over a million and a half views on an episode that was published about six months ago.

That's really a lot.

In three months, it was at one and a half million views which, which is massive.

And it's probably would make it one of the most watched like English podcast episodes on YouTube.

And it's all just two AI characters having what's clearly a fake conversation with each other.

Do you think it's that clear?

Because I looked through all the comments, or as many as I could, under the video that you shared.

And it's like a 26 -minute long video.

It sounds exactly like the AI voices that are generated by Google's Microsoft LM.

The deep dive function, where you put in a text and it generates a roughly 10 -minute conversation between two guests where they talk about the thing that you've put up.

It turns it into a podcast episode.

It sounds exactly like that.

So it's probably generated with the same software, maybe Gemini or something.

I looked through the comments.

I did not see one single comment saying that this is fake, this is AI, nothing.

Everyone is just saying, I love the way you two talk.

This is really useful.

People genuinely really loved it and either a couldn't tell that it was AI generated or B just didn't care?

Yeah, I think it's the first one, unfortunately, but it's, that the people sort of, that it seems so believable.

And I think if you are, maybe if you are at a lower level of English, you know, you're just getting started with your English journey, you might not necessarily recognize it.

I think any native speaker would immediately recognise it, and it's also slightly suspicious that there's no people in the video.

Yeah, it's always the same thing.

It's always an AI -generated image, which is the video.

It's not two actual humans interacting.

It's an AI -generated image of two, really either two very attractive -looking characters talking to each other across a desk, sort of alarmingly good looking.

Or it's two cartoon characters with microphones in a nice happy environment with maybe a sound wave, which is moving, and a text transcript appearing on the screen.

But it's always the same kind of generic American accents, a male and a female.

Um, yeah, but I mean, do you think that if all those people knew that it was AI, would they care?

I'm not sure maybe if, if, uh, if people are listening to this and they have been shown, because I imagine people watching, um, at LEP on, on YouTube presumably have also been shown, you know at first but recommended these kinds of videos and perhaps you've clicked on them and thought hmm that's AI I'm not gonna watch that so maybe yeah add a comment and let Luke know about whether you have tried these or not.

And it seems to me because also the the images they use are pretty clearly AI you don't have to be you know it doesn't matter what language you speak for you to be able to recognize that that is an AI image.

I would think so. I don't know.

But it does seem to me a shame because there's so many people, real humans creating human native contents out there that is much more like, by definition, much more likely to be the kind of thing that you will encounter when talking to a native speaker, yet people are still choosing to learn from AI generated texts, whether they know it or not.

I think probably in the scheme of things, if we calculated all of the listens or views across all learning English channels, and we're just talking about YouTube, YouTube, across all the learning English channels on YouTube.

We worked out the percentage of those views that were from human generated content, and the stuff from AI generated content.

At the moment, I think the human generated content is way more, you know, a channel like English with Lucy, or, you know, one of the other more successful YouTube channels for English.

I think there are loads of channels like that and loads of views on those channels.

It's rare for me to find an AI channel that is similarly successful.

This one that you showed me is the first one I've ever seen.

But I wonder how that will go in the future.

And it's really like the success of these channels is really just based on two factors.

One of them is, yeah, like I said, whether people will actually be able to identify if it's AI or not.

And if for the people listening, if, as far as they're concerned, if it's just English being spoken, then fine.

So that's the one, will they identify it and second factor, will they actually care if it's AI or not?

And whenever I talk about this and this topic does come up quite a lot on my podcast. And I always say the same thing, which is that I reckon eventually people will not care and people always rush to my defence, which is lovely.

And they say, don't worry, Luke.

We'll always choose you or human teachers over AI.

You know, if it's a choice.

And I say, I think to myself that's very nice, but I think that maybe one day you'll change your mind because one day something will happen.

Some content will arrive that will make you go, Wow, that actually that's brilliant.

I love that. And you might choose to watch it or listen to it.

And that sense of there being a negative stigma, you know, attached to it will be gone because everyone else is watching it.

It like this YouTube channel is absolutely massive.

And when you watch it, you realize why.

Because actually, it's great.

It's got all it's just like as good as anything that Luke or Lucy or Alistair would come up with, but even better because of whatever the extra factor is, because it's sort of like somehow is able to pinpoint exactly what a human wants in a way that normal humans are not able to do.

When AI is more intelligent than us, and that apparently is only around the corner, it will be able to manipulate us better than even the best marketing content creators, or whatever you want to call them are able to do.

You know what I mean?

I don't know if that's a paranoid, sort of a slightly pessimistic vision of the future.

But, you know, just from what I've read and heard from the people who really know about this stuff, they do say that eventually when it when the AI is more intelligent than us that, you know, to, a lot of us won't stand a chance, you know, it'll, it'll be able to do things that are brilliant and we will gladly accept them, you know, and I guess also just with the example of this like AI generated podcast. Clearly it's tapping into something that people like because the dialogue, you know, there's two different voices.

The dialogue is clearly at a level that makes it relatively easy to understand and because it's, you know, cause it's all AI generators.

There are not arms and eyes and people accidentally go and changing what they were going to say.

So it's probably pretty easy.

It's like it's like reading a text but just listening to it, which is very different to a normal conversation where people kind of think it's a hang on actually this or they go and they make a slight mistake.

Like, like I'm doing now.

So it's easier. So yeah, you say that, but I've listened to some of these things, and they do incorporate those sorts of natural features where someone appears to be making up their sentence as they go along, like the rhythm of it.

Like previously, AI generated audio, like, let's say even text to speech or something, it did sound very robotic and horrible, awful, unlistenable, but they've found ways to replicate.

They've really found a way to analyse natural human speech and replicate it so that it sounds like someone is thinking while they're talking, you know, the way that, like, I'll do a, give a demonstration of it now.

So when you're making a cup of tea, one of the most important things you have to do is first of all, consider whether you're make the tea in a cup, or whether you're gonna make it in a teapot.

That's the difficult choice that you have to make when you're making a cup of tea." The other person goes, exactly, yeah, that's the theme I'm always having to deal with.

In the cup, or in the teapot?

It's one of those, isn't it?

Yes, it is. Let's have a deep dive into what making a cup of tea is really like, from a British perspective.

It's always that kind of thing where they're sort interrupt each other slightly.

You know, it's incredible what they have managed to develop.

Hello, everyone. I'm just interrupting here, because I'd like to demonstrate what the two of us are just talking about by playing you a sample of an AI -generated audio podcast, which I just created using Google's Notebook LM software.

So what I've done is I went into Google's Notebook LM, which is free, anyone can use it, as long as you have a Google account, and I inputted this podcast episode that you're listening to, this one.

I took the transcript from the conversation that you're listening to and I put it into Notebook LM one, and generated what's called the Audio Overview, otherwise known as the Podcast Deep Dive.

Within a few minutes, it produced this 11 -minute recording.

I'm just going to play you the first minute or so, just to give you an idea of what I'm talking about.

Listen to the way that the speech sounds natural and all those little features I was talking about before, about how it sounds like they're thinking while they speak, the way their voices slow down sometimes the way they react to each other, the way they almost interrupt each other.

Certain things are always repeated, like the word, exactly.

There's always someone saying exactly, always, in these sorts of recordings, and other similar things like that.

Anyway, let's play that sample so you can hear what it sounds like.

Welcome to The Deep Dive.

Today we're looking at a, well, a pretty clever way you can get even more out of listening to podcasts for learning English.

We're talking about using AI tools, the chat, GPT, that kind of thing to really boost your progress.

Exactly. Lots of you are already listening to podcasts, which is, well, it's great for immersion.

Yeah, definitely. But what if that listening could be more active, more targeted, and actually feed directly into your speaking and writing?

OK, I see where you're going.

AI is a sort of helper.

Sort of like a study buddy, yeah, one that's listened to the episode right there with you.

And the cool thing is, this doesn't have to be some huge time commitment room.

Not at all. Even just a few minutes with an AI tool after you listen, it can make a real difference.

OK. So that was a sample of Notebook LM's audio overview automatically created podcast deep dive.

And that's exactly the sort of thing that you find now all over the place, especially on YouTube.

All of these AI -generated podcasts, they all sound like that.

And no doubt, they're just going to get more and more realistic as we go forwards.

What did you think of that?

Was that obviously fake?

Did you think that was obviously fake?

Or did it sound pretty realistic?

Did that sound appealing to you?

Is that the sort of thing you would want to listen to?

Now maybe you were thinking, well, it sounded very American.

Like they just sound really, really generically American.

And, you know, Luke, I listened to Luke's English Podcast, because I want British English.

At the moment, notebook LM doesn't do different accents.

It also doesn't do different languages, I should say.

It's just in English, but it's always that generic standard American accent.

They don't do a British English version, but they will eventually.

There's probably other AI that will do that.

But anyway, how did that sound to you?

Maybe it was a bit too American or something, if you're looking for British English.

But regardless of that, is that the sort of thing you could find yourself listening to?

Did it really strike you as being artificially generated, or did it sound like a conversation that you could have continued to listen to, let us know in the comments.

Anyway, back to the conversation.

And clearly it is, it is better than not listening to anything at all.

So, this is not to say to people who are listening to that, stop doing it.

But it's, yeah, it's fascinating to see that this is a choice that people are making, either consciously or, or subconsciously?

Yeah, indeed it is.

And we got to shape up.

Yeah, we're gonna have to compete.

I suppose, no, I was thinking about that this morning as I was walking down the road.

I was thinking, well, what do we have to do?

We have to really maybe lean into the human side of things, and always keep things really natural and make sure that it's really distinctive and make sure that we're including humor and feelings and things like that, which I think that AI struggles to do at the moment.

I've never really been moved by anything that AI has come up with.

I've been disturbed by it, but I've never really felt a genuine sort of emotional reaction to it.

Yeah. I would agree.

I, I'd also say that, you know, your lived experience and your, what you've shared is about your life as a human being.

Is something that, I know AI cannot replicate.

So if someone's listening to an AI generated podcast and they're saying, oh yeah, this one morning When I was just walking back from taking my children to nursery in Paris, like if an AI podcast is saying that it is clearly just invented.

Uh, whereas there are probably people who've been listening to you since episode number 72, what is the cheese, cheese, um, who have, you know, been following you and your life and its developments since then and will, that, that's stuff that cannot be replicated by anyone else, and certainly not by AI.

We will see. We'll just have to see.

But it's certainly a very interesting time.

We could talk about other stuff, like the things I mentioned before about the effect on the academic world and those things, but I think that's probably enough for now.

Alistair, thank you very much for sharing your specific tips and advice on how people can really use AI as a learning tool with a podcast, for example.

I think that's been very interesting.

I'll remind everyone to get the PDF.

The link will be in the description to get it.

There'll be a lot of the prompts that Alistair has mentioned included there, so you can copy -paste them, adapt them, as you see fit, a link to the LEPGP -T, which a bit of fun.

And some other bits and pieces that you'll find on there.

Great. Alistair, just remind us how we can sort of follow your work and get more from you online, your podcast, your website, and stuff like that.

Sure. As I have a website, which is LeonardoEnglish .com, Leonardo, like Leonardo da Vinci.

And I also have a podcast, which is English learning for Curious Minds, available on all respectful podcast apps and also now available on YouTube.

Okay great. I want to also say, oh, and with one more thing which is why people shouldn't blindly trust anything that AI tells them to do.

As part of thinking about what we should talk about in this episode, Luke, I had some fun talking to Jack GPT about our conversation and I'll tell you what chatgpt said I should end our conversation with.

I'm going to read this out word for word which was probably to explain why you shouldn't do it.

I should end with try these tricks with LEP folks.

Drop your chatgpt results in the comments.

Let's turbocharge your English together.

It just sounds so generic, right?

It does sound quite generic.

it's sort of cheesy.

It's not. That's not the way you would say, you wouldn't say let's turbo charge our English.

Don't forget to smash the like button, folks.

It's not the way that you would have put it.

It's not, it's not really, it's not me.

I think I'll just say thank you very much for having me, Luke.

And, goodbye. All right.

Brilliant. Thank you.

Thank you, Alistair.

Have a lovely day. You too.

Bye. That is basically the end of the episode.

Thank you again to Alistair for taking the time to prepare so much for this episode and to share it with us on the podcast today.

Yeah, just like Alistair said, you can get the PDF.

You'll find a link in the description in the show notes.

You'll find a PDF. And on that PDF is all the stuff you need, really, all of those prompts that Alistair talked about, all the ones that he showed us, plus loads of other ones that we didn't have time to demonstrate.

But you'll find loads of prompts on his PDF.

Lots of stuff that you can use.

You can copy those things and paste them into chatGPT and use them.

You can adapt them as you see fit.

Use them as a starting point.

and yeah, try it all out, okay?

Try to see if those prompts will help you.

Try using ChatGPT or, indeed, other forms of AI, other forms of generative AI, to help work on your English.

And push it further, you know?

Try to push it further and try new ways, more sophisticated ways of using it as a sort of learning assistant or learning tool.

just use ChatGPT to do all the work for you.

I mean, don't just say, I want to write a comment in reply to this message.

Can you do it please?

And then you just copy paste what it does.

You know, don't use it as a replacement for your own ability to use English.

Don't use it as a crutch because then your English will not actually develop properly.

You won't develop those true genuine English skills, but you can use ChatGPT to really help you build your skills as has been demonstrated in the episode.

So try stuff out, try different things and let us know.

I'd love to know what the results were like for you, what you tried to do, what happened, and what the results were.

And if you've got any specific study techniques or study routines or anything that you have found is useful for you, share those things with us.

Let us know how it goes for you.

That's the end of the episode.

I don't want to talk longer further here.

There's a million things to say, as usual.

But I think that's enough for now.

I look forward to reading your responses and comments.

And, yes, link in the description for the PDF.

Thanks again to Alistair.

Have a lovely afternoon, evening, morning, night, whatever time it is, wherever you are in the world.

And I will speak to you again on the podcast soon.

But for now, it's just time to say goodbye.

Bye. Bye. Bye. Thanks for listening to Luke's English Podcast. For more information, visit teacherluke .co .uk.

This podcast is brought to you by Aura.

Imagine waking up to find your bank account drained, bills for loans you never took out, a warrant for your arrest, all because someone stole your identity.

Hackers aren't waiting.

Why are you? That's why we're thrilled to partner with Aura.

Your personal data is a goldmine for hackers, and Aura helps lock it down.

Aura monitors the dark web, blocks data brokers from selling your information, includes a VPN for private browsing, and a password manager to secure your accounts before criminals break in.

For a limited time, Aura is offering our listeners a 14 -day free trial plus a dark web scan to check if your personal information has been leaked.

All for free at aura .com slash safety.

That's aura .com slash safety to sign up and start protecting yourself and your loved ones.

That's a you are a .com slash safety terms apply.

Check the site for details.

We all have that one friend whose opinion we trust in everything for 63 percent of podcast listeners that friend is their favorite podcast host when a cast podcasters endorse a brand that audience listens and takes action.

So if you want a recommendation that really sticks put your brand in their hands book host read sponsorship today by visiting go .acast .com slash ads.

If you enjoyed this episode of Luke's English podcast, consider signing up for Luke's English podcast premium.

You'll get regular premium episodes with stories, vocabulary, grammar and pronunciation teaching from me and the usual moments of humor and fun.

Plus, with your subscription you will be directly supporting my work and making this whole podcast project possible.

For more information about Luke's English podcast premium go to teachaluke .co .uk slash premium info

[Leveraging AI as a Personalized English Learning Tutor]-[936. How AI can Boost your English with Podcasts ✨🎧 (feat. Alastair Budge)]

📋 Summary

Introduction to AI-Powered Language Learning

The "Active Learning" Framework

Practical AI Exercises

The Ethics and Risks of AI

Finding Content and Future Trends

Conclusion

🎯Key Sentences

📝Key Phrases

📖 Transcript