Discussion keeps the world'sturning.
This isRoundtable. You're listening to Roundtable withmyself, HeyYoung.
I'm joined by Yushun in the studio and Josh Caterlale on the line comingup.
I think AI is just about chatbots and social mediaalgorithms.
Thinkagain. We're diving deep into how artificial intelligence is taking on the climate crisis headon.
While also confronting its own environmentaldownsides, how can we harness the potential of AI to combat climate change without making itworse?
And short video platforms may have reshaped our musicaltastes.
Andappetites. Some argue that pop songs should be no longer than two minutes and thirtyseconds.
Are we losingsomething, gainingsomething, or just evolving with thetimes?
Our voice matters and we would love to hear it in our heart to heartsegment.
Got aquestion? Something you're itching toshare?
Fire up that voice memo and hit us up at eZFM Roundtable atFoxMail.com.
Our podcast listeners can find us at Roundtable China on ApplePodcast.
Now let's move on to the next topic of discussiontoday.
As record breaking temperatures and extreme weather events become the newnormal, could artificial intelligence or AI offer a beacon of hope with its potential to provide innovativesolutions, AI may play a crucial role in mitigating and adapting to global environmentalchanges.
But it's also devouring energy at an alarmingrate.
Guys, what is the role of AI in combating climatechange?
Hmm,actually, youknow, at first I thought AI-aiding environmental protection was just a pluffy and hard-to-achieveconcept.
Youknow, how could natural phenomena be explained withnumbers?
Butlater, after somesearching, I found that AI might indeed be able to contribute to environmental protection because I saw that according to an article from World EconomicForum, the nature of climate change management is actually datamanagement.
As we need data todescribe, understand and predictclimate.
AI excels in these kind of handling vast amounts of data from various sources such as satelliteimagery, weathersensors, and historical climaterecords.
This capability allows AI to generate insights and predictions about climate patterns and facilitating better decision-making for mitigation and adaptationstrategies.
So these areeye-opening, Ithink.
Yeah.Well, same question toyou,Josh.
What do you see as AI's role in combating climatechange?
I think that it plays a huge role when it comes to an issue as complex as climate change that really affectseverything.
And it will affect every single facet of daily life for an individual and society as awhole.
And we know thatAI, what it does in a very simplisticway, it's the thing that it does best is processdata, and then give us answers and hopefully solutions based on data that itprocesses.
And I think that when it comes to climatechange, this is really what people are trying todo.
We're trying to collect dataabout, forexample, temperaturechanges, about the consumption of certainproduct,farming,energy, andsuch.
And AI can really help us to give us answers on what to donext,basically.
It's not necessarily what we can do or what we willdo.
And it might not be the most comfortable thing todo, but it can certainly give us somesolutions.
So I think it plays a hugerole.
And to provide you a more stellarexample, you shouldn't has interviewed WangHan-Yuen, founder ofClimemeind.
And this is a platform that gathers a lot of this data and information and uses AI to turn out some of this very much neededresults.
And let's hear from Wang Han-Yuen how this whole thingworks.
So funnyinformation, in more elegantway, what we describe is the access to knowledge is a power ofproductivity.
We have searchengine, we have librarybooks, we can look for answers and all liketeachers.
The revolution of large-engined computer model in AI istotally,well, it's a tipping point in AI fielditself.
But I think like what we're trying to do is to bring that into the climate everyyear, hopefully we can save more time for professional work on that really importantthing, rather than spending a lot of time to do the entry level like finding materials and then to try that increase the cost ofinnovation.
Yes,actually, we know that AI usage is increasingly focused on analyzing these vast amounts of data and conducting different analysis based on varying climates in differentregions.
And this actually requires a lotof, Ithink, experience andlearning.
And thanks to its powerful learningcapabilities, this is actually something AI is so goodat.
And I think that was also one of the initial ideas people had to let AI handle these lengthy yet essentialtasks.
So ofcourse, the utilization of AI in climate change areais, Ithink, promising as we cansee, it will be used in a lot ofareas.
Right. So aside from crunching the numbers and improveefficiency, we see that AI has been touted to be very good at accurately predicting the future for that specific field orpurpose.
Could you give us some of the examples in thatarea?
Ofcourse. On June18th, the China Meteorological Administration launched three advanced AI-based forecastingsystems, functioning for global medium rangepredictions, fun lay for short-term and now casting andfunction.
Well, sound like my name for global sub-seasonal to seasonalforecasts.
These models integrate cutting edge technologies to enhance accuracy and efficiency in weatherforecasting.
And these systems leverage a mix of physics driven and data driven approaches of offering improvements in predicting weather patterns and extreme eventsglobally.
So I would imagine this would be very usefulfor, ofcourse, everyday peoplebecause, youknow, we can plan our daily activity according tothis, as well as foragriculture.
This would be usefuland, youknow, for cities holding all kinds of eventsor, youknow, long-term planning for many aspects ofsociety.
The weather forecast is actually very important and also considering how the climate sort of a justice mood will younilly, it feels thesedays, being able to predict in a pretty accurate manner is so useful also in getting prepared for disasters assuch.
Yeah, I think that when you talk about theweather, that's really important as well because also what AI can do as well as offer assolutions, it can also give uspredictions,right?
That's what it's quite good atdoing.
We hope that it is anyway rightnow, but I think that when it comes to improving thingslike, youknow, the prediction of weather patterns that can help governments to createpolicy, which actually leads us on to one of the issues because we're talking really here about globalpolicy,right, which we know can be incredibly complicated when it comes to itsimplementation,right?
But eitherway, it can certainly offer us some policy and better guidepolicy.
Forexample, I know that acompany, a massive company such as IBM has something called Green HorizonsInitiative, which uses AI to protect air pollution levels and has helped cities like Beijing actually to reduce air pollutionlevels.
And there's also AI-driven early warning systemsfor, as you mentionedbefore, naturaldisasters.
And I also found another example that Google created an AI Flood Forecasting model that provides pretty accurate predictions up to 48 hours inadvance.
There's a multitude of other examples in thisregard.
So I think that you made a great point that it can also help with predicting a lot ofthings.
Butagain, that also leads us on to an issue which is implementing such policy on a globallevel.
Yes, I'd say the devil yes is indeed in thedetails, but it's also in the implementation of a lot of these grand ideas and sometimes really goodtechnology, but can it really be maximized its effect comes toimplementation.
And also AI has been used in environmental monitoring and this is one crucial step before we can talk about disaster control and all those different other aspects ofthings.
Absolutely.Actually, nowadays we can see that the use of satellites is very commonglobally, but with the information that we get from thesesatellites, we can even better useit.
Forexample, mappingdeforestation, Edinburgh based company space intelligence operates in over 30 countries and has utilized satellite data to map over 1 million hectares of landglobally.
Their technology can remotely assess metrics like deforestation rates and carbon storage withinforests, which will be lengthy and time consuming for human to identify icebergs amid the white of clouds and seaice.
Andalso, since we mentioned something aboutagriculture, I think the utilization of AI in this area is also very important in the way that they can maybe predict the yield in specific areas by analyzing also the satellite images and tracking cropgrowth.
And with the accurate forecast of a lot of these croppedproduction, these governments and organizations can make better plants or farmers can have better plants on planning different areas for cultivating different types ofcrops.
Right. And when we talk aboutAI, I think for mostfolks, it sounds kind of like an ephemeralthing.
Where does ithappen? You don'tknow.
But infact, AI is run on computers and a whole bunch ofcomputers.
It's got real life environmentalcosts, which we oftenforget.
I think it's an important point that youmake,Hayoung, and it's important to talk about the energy consumption that AIuses,right?
And I think that most AI computer systems are highly energyintensive,actually.
And there's a lot of data centers powering AI computer companies that consume vast amountsof,obviously,electricity.
I found one study that estimates that training a single AI model can emit as much carbon as five cars over theirlifetime.
I think thatwe, as asociety, have a decision about where we go withthis.
And we can either decide to make it more energy saving ornot.
Anyway, there'salso, ofcourse, off the back ofthat, the carbon footprint that AIhas.
I found oneexample, apparentlytraining, chatGPT-3, estimated to have emitted 552 metric tons ofCO2, which is the equivalent to the carbon output of 120 US homes in ayear.
Nowagain, that's quite alot, but it depends where we go withthis,right?
Does that mean that in the future we're able to make more US homes more carbonneutral?
If that's thecase, then there's a balancethere,right?
That's quiteachievable.
So I think this is where we'reat, in myopinion.
Right.Well, I'm not sure if I share thatoptimism.
I hope Ido. I hope the world can go thatway.
And I will also add more figures to what Josh have alreadysaid, the International Energy Agency or IEAestimated.
This was from earlier thisyear.
Electricity consumption associated with datacenters, cryptocurrency and artificial intelligence addedtogether, represented almost2% of global energy demand in2022.
And that demand for these uses could double by 2026 if we're going on the same path at thisspeed, which would make it roughly equal to the amount of electricity used by the entire country ofJapan.
Andalso, this is kind of adding insult toinjury.
Not only is electricity such crucial to powering AI and all these data centers and GPUs needcooling, just like yoursmartphone.
Youknow, it heats up after you looking at videos for twohours.
Andtherefore, for these data centers and GPUs to cooldown, it'll need to use more and morewater.
And just think about all the chips that go into all this AI powerdevices.
That's a lot of energy consumption to make thesedevices, to train the AI models out there and also for us to use thetechnology.
Sowell, given allthat, do you feel that maybewe're, although AI one could argue is still at its nest and stage for consumeruse, but don't you feel that we need to make this important decision asindividuals, asusers, before it's toolate?
I wonder if you would still be an avid user after listening to today'sshow, because with what's been talked about outthere, it seems like AI's solution to so manythings.
And it could be something as small as what's my recipe for the big party that I'm going to host for a couple of friends over theweekend, or how am I going to plan my trip nextmonth?
All of these things you can consult with generativeAI.
But do you feel a little bit differentlynow, like hearing thatactually, although you don't feelit, but somewhere outthere, there's real life environmental cost that is spent on our rather frivolous demands fromAI,maybe.
Maybe I'm just being too ignorant on the consumption of all of these chips and AImodels.
But what I was thinking is that that is probably the process and the procedure of the development oftechnology.
And all ofthese, Ithink, inventions from the history we cansee, there are some of thecost.
What I was thinking is that now many major tech companies are promoting that theirretail,stores, or office buildings have achieved zero carbonproduction.
This is kindof, Ithink, forward thinkingmeasure, not only for the future of thecompany, because this is obviously a great promotionalfocus, but also for the future of the entireindustry, because sustainable energy might be one of the best ways to combine technology companies with environmentalprotection.
So I think with the flow of developing advancedtechnology, Ithink, maybe we need to come up with better solutions instead of just stop usingit.
Well, I think that off the back of that one thing that I think is particularly worrying or maybe even more worrying than maybe the energy consumption is the risk of the loss of human skill and dependency onAI.
I think that for me personally rightnow, and just as you should havesaid, I may also be a little bitnaive, I think we're all naive to some degree about what the future holds withAI,right?
Because we just don'tknow.
But I think that one risk that I see I personally worry about is the dependency onit.
And when it comes to things like youmentioned, heryoung, we can talk about hugethings, like combat and climatechange.
We can also talk about small things like how do I plan my evening with myfriends?
Oh, I'll just ask generative AI to helpme,right?
And I think that the more we rely on it for those smallthings, the more that we can maybe just lose our humanexpertise.
And I don't know where that's going to leave us asindividuals, as a society in theend, especially when we start using AI to deal withAI, which is what we're already starting todo.
Imean, it sounds like a pretty scary rabbit hole to go down forme.
Yeah.Well, the thingis, these days with all techcompanies, they're engaged in a pretty obvious AI armsrace.
And then they've also trained and built the AI chatbots or virtual assistants in such a friendly way and so accessible aswell.
Likenow, a lot of this AI chatbot function is built into various apps and even some of the like drawing apps have ittoo.
So it's almostlike, youknow, you don't even think aboutit.
Okay, let's just get the chatbot out and then help me lay the laborious and boringgroundwork.
And then I'll just build on top ofthat.
But the more we're used tothat,yes, I think it has something it will have some kind ofinfluence, maybe even to our cognitivefunctions, but also to the skills that we're able to acquire andutilize, maybe to live in thisworld.
Butalso, I just feel that the environmental cost is so behind on the back burner that we don't realize we don't think about and then I think one of you maybe have touched upon this aswell, but I'll use a differentanalogy.
I just really just like this idea of fighting fire withfire.
And in this technological drivenworld, I feel a lot of the times our solution to issues brought by technology is we'll just use moretechnology.
And then that just pushes me down that maybe a different rabbit hole as the one that Josh is currentlyin.
So what is the future hold when it comes to utilizing AI and let's faceit, this huge issue of climate change or some call a climatecrisis.
It kind of depends on where you liveat.
And if you happen to live in a zone where this is happening in front of youreyes, you'll realize how big of a deal itis.
And I think for those folks like you and I sitting in this studio with the very comfortable AC or air conditioningon, we are extremely lucky that we don't see theflooding, theearthquakes, the melting of ice caps as such in front of our veryeyes.
But you knowwhat, it's happening not that faraway.
So that's enough of myrant.
In youropinion, what do you see the future holds for AI and its use in fighting climatechange?
There's a place you can reachus.
EZ FM around table atFoxmail.com.
Shoot us an email or even better a voice memo and share your opinion withus.
Coming up has pop music just turned short and sweet or our listeners being short-changed we take a look at why pop songs are getting so much shorter thesedays.
Staytuned. Looking forpassion?
How about fierydebate?
Wanna hear about current events in China from differentperspectives?
Then tune in to roundtable.
Where East meets West and understanding is thegoal.
Is the hour of round table formyself,young?
I'm joined by you Shun in the studio and Josh Caterl on theline.
Music as an integral part of human culture has always evolved in tandem with technological advancements from vinyl records to cassette tapes from CDs toMP3s.
The most recent and arguably the most impactful shift in music consumption has been the rise of streamingservices.
In the pastdecade, streaming services such asSpotify, Apple Music and domestic counterparts such as NetEase and QQ Music have revolutionized how we listen to and discovermusic.
Then came the do-in and tic-tocks of the world and short video platforms have altered the way we consumemusic.
And have you noticed that songs are simply getting shorter andshorter?
Is it because that music is reaching us through these platforms and therefore it has produced suchresults?
What do you see ishappening?
Actually we can see with the rise of all these streaming and short videoplatforms, there is now increasing discussions about the length ofmusic.
Some argue that longer songs can explore deeper emotions andthemes.
While others believe that it iscorrect, if you completely just ignore the storytelling aspect ofmusic.
I saw some young singers have expressed online that they believe we don't need to repeatverses, we don't need bridges and we don't need a lengthyoutro.
Of course the length of a song depends on the creator'sintention, but this view of creating shorter music seems to reflect the mindset of many younger creators and the fact is our music is gettingshorter.
According to the data of Billboard since1990, the average length of a song on the Billboard Hot 100 has decreased from over 4 minutes to around3, regardless ofgenre.
And what abouttoday? Now we're looking at songs that are less than 2minutes?
Yes, the trend of shorter music is evident on the music charge aswell, according to media report at the 2024 GrammyAwards, 28 of the 144 nominated songs come in under 3minutes.
Morespecifically, Harry Styles as it was is 2 minutes and 48 seconds long and Sizzas Kill Bill is only 2 minutes and 33seconds.
Globally, K-pop and New Jean is 1 minute 48seconds.
Yeah,well. Ohyeah. And withoutexception, these songs have garnered billions of streams as it was has over 3.2 billion streams and Kill Bill has over 1.8 billionstreams.
Well, I think that really what we're talking about here is society as awhole, very rapid cognitive decline and inability to have critical thinking and retain attention for more than 2minutes.
I don't think that this is only withmusic.
I think this goes for film and also for things like reading and just generally becoming addicted to the dopamine hit that you get when you see these shortvideos.
These are getting shorter and shorter and it's affecting the music industry aswell.
I think that we already know that it's affected us with things like social media andfilm, but because music for a longtime, pop music especially has always been about 3 minuteslong.
Andactually, if you go back to maybe one of the first modern popbands, which was theBeatles, actually a lot of the Beatles biggest hits are less than 3 minuteslong,right?
So what's the differencehere,right?
I think that there's also some differences not just in thelength, but also in the content of themusic.
If you look at a lot of these songs that are very ticktock, youknow, friendly and blow up on apps like ticktock, they're also very simple in how theysound.
They're very repetitive and they'll focus on a part of their music that is maybe easy to sing along to regardless of what your first languageis,right?
So maybe English isn't your firstlanguage, forexample, but you're able to easily singit.
Also, it'sdanceable. So maybe it only has to have say 15 seconds of catchy music and that'senough,actually.
And so a lot of musicians are focusing in onthis.
And I do think that it is having a negative effect on the musicindustry.
Ido. I think that it's important to have really well thought out pieces of music and art ingeneral.
I think that that's good forsociety, it's good for cultural and artisticappreciation.
And I think there are psychological benefits tothat.
So,but, youknow, at the sametime, there are somepros.
I think that this offersartists'opportunities,right?
Everything can be seen as an opportunity and we have seen with apps like tick tock artists who maybe aren't sofamous, suddenly seeing their song blow up maybe six or seven years later because maybe an influencer catches onto it creates a dance toit.
And then the song becomes bigger and then it becomesfamous.
So there are winners here aswell.
Yeah,absolutely. With the appearance of all of these short videoplatforms, there are even some artists that design or create music solely for short videoplatforms.
Ofcourse, the impact of streaming and clicks plays a significant role in I think thechanges.
Shorter songs are more likely to be replayed on theseplatforms, given them an advantage in terms of traffic andclicks.
Andalso, the streaming algorithms may be more favorable to shorter songs because with the shortersongs, you may click it moretimes.
And in thisway, they can have moreclicks.
And then most music platforms incorporate skip rates into their recommendationalgorithms.
And these songs may be recommended to a largeraudience.
Right. This is another example where the algorithms of the platforms sort of dictate the coreforce.
I found one smallfact, which I find to be reallyinteresting, apparently on streaming platforms for it to recommend newmusic.
It looks into the rate whether this song has been listened to from beginning toend.
Andtherefore, if you are a super long song like NovemberRain, which goes on for famously 12 minutes or something likethat, who has the time to sit through that amazing guitar solo thesedays?
Anyhow, so artists aren't writing these kinds of songsanymore.
And then they want their song to be listened to from beginning to end so it can be recommended to morepeople.
And then it changes the way possibly how musicians are writing their music and presenting theirmusic.
Do you think maybe it'spossible, let'ssay, if with these short video platforms or newplatforms, maybe we get a quick taste of the music and then maybe we can go back to the originalform, the way that it's supposed to be celebrated by fans andlisteners.
Maybe, or maybe I'm just asking too much ofpeople.
People just want quickfix, quick entertainment andrelaxation,meal, and that'sit.
Or they don't even sit through themeal.
They just want theappetizer.
Am I too negative about this pessimistic aboutthis?
I think that there are some positives to take away from this aswell.
Ofcourse, we can learn to appreciate longer form music and I think that there'll always be a market forit.
And ofcourse, there's a lot of convenience to this aswell.
And it allows musicians to get their music out there when otherwise they might find that difficult to doso.
Allright.Well, that brings us to the end of today's roundtable.
Thankyou. You should and Josh Cotterall for joining thediscussion.
