The US anti-trucks and consumer protection watchdog has recently launched a probing to what it calls surveillance pricing.
There are seeking information on the use of artificial intelligence to set different prices to the same product for different consumers.
It is as similar to what Chinese people had experienced and what the Chinese legislators have tried to combat.
Hello, I'm Limeing and this is Roundtable.
Today with Li Yi and Steve Hathley we find out whether actions are required and if so are they sufficient to prevent discrimination and unfair competition in the marketplace.
So the FTC, the Federal Trade Commission, they are trying to do something they're investigating into issues of companies trying to collect information about consumers in the US and using the data to determine different prices for different customers.
They are indeed and this isn't a new issue per se but how they're doing it seems to be the new thing.
How this story came about in the news this time is kind of interesting.
So here's what happened.
There was a reporter from CNN her name was Elizabeth Buckwald and she got a text message from a friend saying oh Starbucks has a buy one get one free promotion open up your app on your phone and you'll see it there.
Elizabeth opened up the app on her phone but there was no buy one get one free promotion there so she thought huh what's going on.
So what she did was now remember this is a CNN reporter right so put on her reporter hat and went to work.
She contacted Sheka Gine who is a lead partner in the North American consumer and retail division at the consultancy firm called Simon Kutcher and according to Sheka it's likely that Starbucks used artificial intelligence to determine that her friend if offered a promotion would make a purchase.
Here she would otherwise not have made.
Sheka also said it's impossible to say exactly how Starbucks systems made that correct judgment about Elizabeth the reporter though her past shopping behavior and the shopping behavior of similar customers were almost certainly factors in determining that Elizabeth's friend would not have made a purchase without the coupon but Elizabeth would have made
that same purchase even without the coupon and that's why Elizabeth never got that coupon.
So the artificial intelligence technology was at play here determining through one way or another that we couldn't tell because we are not technology gurus here and even the company itself couldn't tell whether or not this is well entirely logical or something.
Could it or wouldn't Elizabeth called Starbucks contacted them and asked so what's going on here with the AI model.
Their AI model is called Deep Brew and a spokesperson from the company they wouldn't answer that question specifically but the spokesperson did confirm AI is powering the individualized offers that it sends to its customers.
But well correct me if I'm wrong but AI of course employees some fancy technologies like neural networking and deep learning but still there are algorithms right.
They are. They are algorithms.
And whoever coded or wrote the algorithm should be able to tell how the system works.
Well that's true. It's whether they were willing to share that information or not.
Whether they're ready to share the indexes or the factors that are at a play here.
Somebody's income where they live and shopping experiences or perhaps even raised but I'm sure they wouldn't come find on that.
They wouldn't come find on anything according to that particular spokesperson.
Again it's not a brand new issue and by the way this isn't only Starbucks.
This is just the example in the story.
Plenty of companies are increasingly leveraging customer data.
They get that data from their loyalty programs a lot of the time and then they do this in coordination with these machine learning models to customize prices of certain goods and certain services based on an individual's willingness to pay.
And the end goal going back to Shikagin again was to quote get you to buy more items by the same items again or to spend more on the same items.
I think here in China we also have like similar debate surrounding this similar topic.
I'm not really sure if you have heard of this phrase here in China called tashu ju shashou meaning that a pricing system that majorly targeting frequent or familiar users or consumers involving big data and that's becoming an issue here in China.
As you said it's not really something new.
I think operators business operators they have been practicing this pricing system since a long time ago but this time is different because AI is involved.
The meantime big data algorithm has I mean have been playing a role in terms of establishing this sort of personalized pricing system.
And here in China I just want to give some background information about this pricing system as well because I think we have been observed this existence of this pricing system.
Yes since a long time ago but it has only like sparked public debate since 2018 when the phrase tashu ju shashou first really being recognized publicly and also sparked public debates of course.
And let me just give you one example to show how widely the discussion about tashu ju shashou or big data driven or enabled pricing system or price discrimination was that year the phrase tashu ju shashou become one of the top 10 trendy new keywords about social issues in that year and that's according to a real rank by national statistics statistics authorities
and it also called official attention back then it also ranked among the top 10 consumer rights infringement cases in 2018 and that's also an official list jointly released by the Ministry of Public Security and the China Consumers Association and other bodies and also very recently I think here in China we also have a scene this discussion about big data or AI
driven price discrimination because summer vacation is going on here and a lot of people are actually driving across the country and when they are booking tickets either I are playing tickets or accommodation hotel tickets they surprisingly find that different people using different mobile phones can actually enjoy or have different prices depending on the brand of phone
that you own yes that that is also the case so it's not really targeting different customers but also different devices different platforms and all this are being included in this issue different devices is different customers and it's a different demographic isn't it there was an example from 2012 orbit so orbits is a travel website in the United States and in 2012
they were directing Mac users to higher priced hotels why well they had determined not only they had determined but others as well had a sense determined yeah Mac users had more purchasing power well then the Wall Street Journal investigated them and they stopped that practice so yeah different hand sets different tablets different computers how do we feel about this is this discriminatory
I think it obviously is but how much does it actually bother you I think for customers we suddenly feel betrayed because well intuitively speaking we tend to believe that businesses are likely to roll out royalty programs that the more you use their service the more preferential treatments you are supposed to get from them that was the intuitive way of looking
at things but then upon learning that they're actually using that data to give you a disadvantage in terms of pricing then that that suddenly gives a sense of betrayal you should be able to give me more discount on top of that you're charging me more for using your service yeah I mean think about this there's a toothbrush for sale and Lyming needs a new toothbrush and I
need a new toothbrush why does Lyming get a coupon for that toothbrush and I don't because I don't brush I don't brush was a day well needs to toothbrush that much people are upset about this and and this is why and you talked about this in the opening Lyming this is why the FTC has gotten involved the FTC is the Federal Trade Commission and if you don't know what they do
they enforce a variety of anti-trust and consumer protection laws that affect virtually every area of commerce and they're referring to this particular case as surveillance pricing and they want to know what's going on FTC Commission sent orders to eight different companies master card is one revionics is one and the list goes on and they want those companies
to provide information on how they allegedly offer surveillance pricing and services that quote incorporate data about consumers characteristics and behavior they're going to study this the FTC is I don't think you're going to find any results coming out in the next week or two or three or four it's going to take a while but the fact that the FTC is on this now is a sign
that the government is looking to fix this problem too I think also similar issue here in China because after I mean this widely debate about this big data driven price discrimination I think authorities are also releasing greater actions against this system and including releasing sort of regulations because because when consumers are complaining about this pricing
system one challenge for them is that they find it's quite different I mean difficult to define what is this personalized pricing I mean how do you define pricing discrimination that could be a be a challenge and I think it's somehow being answered by Chinese regulations saying that this sort of behavior I mean given giving different prices based on consumers
different backgrounds age or whatever criteria could possibly constitute abuse of market monopoly so that's also an issue being discussed here because obviously when you see the rise of use of digital tools in a lot of online platforms I mean a majority of life services are being diverted to online platforms especially here in China that's also why we see this rise
of discussion about this pricing system because everything is digitally everything is happening digitally so when the word monopoly comes into place then we realize that perhaps the size of the company that is that we were talking about matters for instance a bigger company forings tech giant that has millions if not billions of users might certainly have access
to much more personal data of their users and by that advantage they're able to build sort of a trench around around their businesses I mean they can either sell that data to whoever wants that or they can develop a particular pricing systems or even products tell it at their consumers whereas perhaps a smaller upstart might not even have that advantage well the smaller
upstarts or the smaller companies still have access to this kind of assistance if you will one of the companies that I mentioned before revionics that the FTC is looking at the vice president of communications for that company her name is Kristen Miller and she said that the company does not in any way conduct operations related to the surveillance of consumers
we are confident that the FTC will affirm the benefits of revionics is AI price optimization software now what revionics does is they will work with a company and they'll help them to come up with pricing for their products by giving them different models one of the examples that I found was from a farming equipment company called tractor supply co and the CEO of that company
her name is Mary Wynn Pilkington and said she said we work with a company like revionics to help us identify our customers our customer base and provide them a better quality product but we don't ever want to alienate them we want to attract customers we want to retain customers but we're not trying to take advantage of customers but even with this supply of such service
in the market that conveys an extra cost for upstart companies whereas perhaps a bigger company like a tech giants an Amazon or something they naturally would have access to that data they don't have to incur that cost in order to base the decision upon them so that's the advantage of you a bigger company I suppose right so which is why the the authorities are looking
at the issue from market fan is competition that kind of approach the other way to look at this is whether it is fair for customers consumers as well because on depending on how the mechanism works most likely when we say partial to social we're describing a phenomenon where more frequent users are being charged more but in reality but really in essence is that is the logic
that the more you need a product the more badly you need it the more likely you're you're going to be charged more so by that standard comfortable rich well-off families they probably have more than what they need and they have less of an incentive to spend more on the same product essential products like food and then perhaps stationary and essentials at work and in school
whereas poor families who probably don't already have them they are badly in need of such products food or stationary or books and then in that case they are charged more and so then you will have a situation where the wealthy spent less whereas the poor spent more and it's a very bad situation where the poor gets poorer or there's another possibility that we're seeing
already happening here because in theory I think as you said the big data or AI will perceive that richer families were richer consumers would be willing to pay more and that's why they are also charging more for those consumers and that's already happening because when you look at those accommodation or hotel booking platforms you know they've got their own membership
system right so if you are a consumer that is I mean frequently using their services so there's possibility that you are willing to pay for their membership so that you can get discounts but in reality some consumers are surprisingly finding out that even though there are members for those platforms they are even charged more than regular consumers because the system
somehow perceive that since you are already a member of our services you're much more willing to pay for yeah or you can afford it so we're going to charge you more that doesn't work no so the result is unpredictable again but the the other issue is the issue of resources a waste of resources because at least we can predict that those who need a badly they're going to
be charged more whereas those who don't need a badly will be given a discount and therefore an incentive to buy something so you end up having more consumers buying stuff that don't they don't necessarily need and this is to me consumerism and really a waste of resources it suddenly helps the economy but I'm not sure it really helps the humanity and now I'm planning
the end of this story is this very different I'll play devil's advocate just for a moment do you remember and maybe you still get them now if you go to your apartment building or wherever you live and you see something stuck on your door it's an advertisement for the local supermarket and they're offering coupons for a discount on I don't know rice or vegetables or what have you
there's a sale on those items where if that same grocery store chain exists in an ultra-rich neighborhood as well for example they don't put any discount flyers on people's doors there is this really what we're talking about right now is it any different than that example they're targeting a community as opposed to an individual offering discount prices trying
to drum up sales in that community whereas in the other community they don't need to do that because they are already they're already shopping at their store is this different I think it's really different because if you compare I mean physical business with online business so it really involves different costs for business operators I mean the example you mentioned
grocery in say communities with residents with perhaps lower income and compared to richer communities I mean cost rent for groceries are also different but one comes to online businesses I don't really think a different consumer groups matter so much for business operators in terms of their cost which is why I think it's really not reasonable for them to have different
pricing system the economists and businesses tend to look at issues in a very simplistic way I mean if a business is willing to sacrifice a portion of their revenue or income by handing out discounts and then by an in-sou doing by expanding the consumer base if the calculation works in their favor then by all means they should be able to do that but the real world is much
more complicated like Steve mentioned earlier a zip code isn't just a zip code a zip code is a community inhabited by you know a group of people it's a demographic yeah it's a demographic and then they could be race involved the incoming valve class everything is involved and when you've factored them all in then there's the issue of political correctness discrimination and then
so it's the real world is really complicated and which is why we're having a discussion about this whether this is discrimination whether authorities should do something about it and also you know one challenge I mean one concern for me is also that with the rise of usage of digital tools and also everything is being produced or being provided through an online
form consumers are finding it increasingly difficult to escape from this kind of system say if you are consumers think you are experiencing price discrimination then theory if you are offered an alternative option that you can escape or you simply say I don't really want to choose this service I just want to choose other services that is also fair but the thing
is that with the prevalence of this digitalized tools I think somehow I mean if you don't like this platform you think you're being swendled then you already have this possibility to you know turn to another platform because that one is also using the same pressing system which is old during by AI and big data in that case you're stuck with the good old fashion way
of going to a brick motor store off the grid or get a second phone and only make purchases through one phone and get the discounts on the second phone it only take them about one two days a week to get to confirm figure out that you are who you are because the location where you place the order is the same is it not terrifying that AI can now predict you know not only how much
you're willing to pay for a product they can also predict when you're going to buy that product based on your previous buying trends they can know where you are likely to make that next purchase as well where when I mean this is this is kind of scary and for me the issue is transparency right that reporter from CNN called Starbucks looking for confirmation that AI was doing
this and they said well AI is driving the but we can't say it's a black box I couldn't tell we can't say any more than that whereas putting a flyer on your door that says carrots are minus 25% until Thursday that's very transparent people everybody knows what's going on here nobody really seems to know about this and that's why the FTC is looking at this to make sure that everything
is on the up and up I know and the other issue perhaps a spill over effect and that is the differential pricing system provides an incentive for companies to collect more data about consumers and when it becomes a common practice to collect more and more personal data the issue arises about way how are we supposed to keep the data safe because when everybody's collecting
data then when data is stored everywhere then it's the risk of well cracking hacking and falling into the wrong hands sure yeah I don't know I remember the days when all of this just started you know you remember you would search for a new pair of jeans on one website and then you would go to your social media site and all of a sudden the advertisement for that exact same
pair of jeans would be in your feed and I remember having the conversation with my friends and my friends at the time saying like oh I don't like this at all I don't like the direction where this is going and my response at the time was I actually don't mind the reminder I was like oh yeah right I'm into buy those those jeans and then I would go ahead and buy them obviously
their structure worked perfectly and it happens with airplane tickets doesn't it you search on one site it'll pop up on another but now it seems like it's an even more extreme example it is right and it now lining here's where we are today in 2024 if the FTC doesn't step in if if authorities don't step in in China and try to regulate this then where are we going to be in 2027
or 2030 well thank perhaps by 2027 China will have a demanded version of the law against unfair competition because a discussion about in fact reference to to to differential pricing was already included in the 2022 draft amendment and and so far I haven't found any update about Wayne is going to be passed by I'm sure by 2027 you should have something and last question that is we
are taking the word at the best value from the company saying that AI is involved and we don't know how exactly AI works and that makes the question should authorities step out their effort in supervising these these situations because if you look at the example of Boeing which is rather high tech and they somehow are so high tech that the authorities are don't necessarily
have the capability to supervise their business to make sure that the planes are safe so the system of checks and balances has to have the ability of checking and balance you know so and but that also the challenge is that whoever has ability to supervise these companies they can easily make a lot of money if they go into the private sector so that is something that's
kind of scary that we need to think about we'll be talking about this in a years time I'm sure yeah
