Latest / Masters in Business / At The Money: Fan Favorite - Algorithmic Harm

Transcript

At The Money: Fan Favorite - Algorithmic Harm · Masters in Business

• 0:02
  Speaker 1: Bloomberg Audio Studios, podcasts, radio news.
• 0:26
  Speaker 2: Algorithms are everywhere. They determine the price you pay for
• 0:30
  Speaker 2: your Uber, what gets fed to you on TikTok and Instagram,
• 0:35
  Speaker 2: and even the prices you pay in the supermarket. Is
• 0:40
  Speaker 2: all of this algorithmic impact helping or harming people? To
• 0:45
  Speaker 2: answer that question, let's bring in Cass Sunstein. He is
• 0:49
  Speaker 2: the author of a new book, Algorithmic Harm, Protecting People
• 0:53
  Speaker 2: in the Age of Artificial Intelligence, co written with ornbar Gil.
• 0:58
  Speaker 2: Cass is also a professor at Harvard Law School and
• 1:02
  Speaker 2: is perhaps best known for his books on Star Wars
• 1:06
  Speaker 2: and co authoring Nudge with Nobel Laureate Dick Taylor. So, pass,
• 1:12
  Speaker 2: let's just jump right into this and start by defining
• 1:16
  Speaker 2: what is algorithmic harm?
• 1:20
  Speaker 3: Okay, So let's use Star Wars. So let's say the
• 1:23
  Speaker 3: Jedi Knights use algorithms, and they give people things that
• 1:28
  Speaker 3: fit with their tastes and interests and information, and people
• 1:32
  Speaker 3: get If they're interested in books on behavioral economics, that's
• 1:35
  Speaker 3: what they get at a price that suits them. If
• 1:38
  Speaker 3: they're interested in a book on Star Wars, that's what
• 1:41
  Speaker 3: they get at a price that suits them. The Sith
• 1:44
  Speaker 3: by contrast, take advantage with algorithms of the fact that
• 1:48
  Speaker 3: some consumers lack information and some consumers suffer from behavioral biases.
• 1:54
  Speaker 3: So we're going to focus on consumers first. If people
• 1:57
  Speaker 3: don't know much, let's say about healthcare product, an algorithm
• 2:01
  Speaker 3: might know that that they're likely not to know much
• 2:04
  Speaker 3: and might say, we have a fantastic baldness cure for you.
• 2:08
  Speaker 3: Here it goes, and people will be duped and exploited.
• 2:12
  Speaker 3: So that's exploitation of absence of information. That's algorithmic harm.
• 2:16
  Speaker 3: If people are super optimistic and they think that some
• 2:19
  Speaker 3: new product is going to last forever, when it tends
• 2:22
  Speaker 3: to break on first usage, then the algorithm can know
• 2:26
  Speaker 3: those are unrealistically optimistic people and exploit their behavioral bias.
• 2:31
  Speaker 2: So I referenced a few obvious areas where algorithms are
• 2:37
  Speaker 2: taking place. Uber pricing is one the books you see
• 2:43
  Speaker 2: on Amazon is algorithmically driven. Clearly, a lot of social
• 2:48
  Speaker 2: media for better or WORSEUS algorithmically driven, and even things
• 2:53
  Speaker 2: like the sort of music you like on Pandora. What
• 2:56
  Speaker 2: are some of the less obvious examples of how algorithms
• 3:02
  Speaker 2: are affecting consumers and regular people? Every day.
• 3:07
  Speaker 3: Okay, So let's start with us straightforward once and then
• 3:10
  Speaker 3: we'll get a little subtle. So straightforwardly, it might be
• 3:13
  Speaker 3: that people are being asked to pay a price that
• 3:18
  Speaker 3: suits their economic situation. So if you have a lot
• 3:21
  Speaker 3: of money, the algorithm knows that maybe the price will
• 3:23
  Speaker 3: be twice as much as it would be if you
• 3:26
  Speaker 3: were less wealthy. That I think is basically okay. It
• 3:31
  Speaker 3: leads to greater efficiency in the system. It's like rich
• 3:34
  Speaker 3: people will pay more for the same product than poor people,
• 3:38
  Speaker 3: and the algorithm is aware of that. So that's not
• 3:41
  Speaker 3: that subtle. But it's important. Also, not that subtle is
• 3:45
  Speaker 3: targeting people based on what's known about their particular tastes
• 3:49
  Speaker 3: and preferences. Let's put wealth to one side, and so
• 3:52
  Speaker 3: it's known that certain people are super interested in dogs,
• 3:55
  Speaker 3: other people are interested in cats. And there we go,
• 3:59
  Speaker 3: and all that is very straightforward's happening. If consumers are
• 4:02
  Speaker 3: sophisticated and knowledgeable, that can be a great thing to
• 4:07
  Speaker 3: make markets work better. If they aren't, it can be
• 4:10
  Speaker 3: a terrible thing to make consumers get manipulated and hurt.
• 4:14
  Speaker 3: Here's something a little more subtle. If an algorithm knows
• 4:18
  Speaker 3: for example, that you like Olivia Rodrigo, and I hope
• 4:22
  Speaker 3: you do, because she's really good, then there are going
• 4:24
  Speaker 3: to be a lot of Olivia or Rodrigo songs that
• 4:27
  Speaker 3: are going to be put into your system. And let's
• 4:30
  Speaker 3: say no one's really like Olivia Rodrigo, but let's suppose
• 4:33
  Speaker 3: there are others who are vaguely like her, and you're
• 4:36
  Speaker 3: going to hear a lot of that. Now, that might
• 4:38
  Speaker 3: seem not like algorithmic harm, that might seem like a
• 4:42
  Speaker 3: triumph of freedom and markets, but it might mean that
• 4:46
  Speaker 3: piece of people's tastes will calcify, and we're going to
• 4:49
  Speaker 3: get very bulkanized culturally with respect to what people see
• 4:54
  Speaker 3: in here. So they're going to be Olivia or Arigo people,
• 4:56
  Speaker 3: and then they're going to be led Zeppelin people, and
• 4:58
  Speaker 3: they're going to be Frank Sinatra people. And there was
• 5:00
  Speaker 3: another singer called Bach. I guess I don't know much
• 5:03
  Speaker 3: about him, but there's Bach, and there would be Bach people.
• 5:06
  Speaker 3: And that's culturally damaging, and it's also damaging for the
• 5:11
  Speaker 3: development of individual tastes and preferences.
• 5:14
  Speaker 2: So let's put this into a little broader context than
• 5:19
  Speaker 2: simply musical tastes and I like all of those, so
• 5:23
  Speaker 2: I haven't become bulkanized yet. But when we look at
• 5:29
  Speaker 2: consumption of news media, when we look at consumption of information,
• 5:33
  Speaker 2: it really seems like the country has self divided itself
• 5:38
  Speaker 2: into these happy little media bubbles that are either far
• 5:42
  Speaker 2: left leaning or far right leaning, which are is kind
• 5:45
  Speaker 2: of weird because I always learned the bulk of the
• 5:48
  Speaker 2: country and the traditional bell curve, most people are somewhere
• 5:53
  Speaker 2: in the middle. Hey, maybe they're center right or center left,
• 5:56
  Speaker 2: but they're not out on the tails. How does the
• 6:00
  Speaker 2: these algorithms affect our consumption of news and information?
• 6:07
  Speaker 3: About fifteen twenty years ago, there is a lot of
• 6:10
  Speaker 3: concern that through individual choices, people would create echo chambers
• 6:16
  Speaker 3: in which they would live. And that's a fair concern
• 6:19
  Speaker 3: and it has created a number of, let's say, challenges
• 6:23
  Speaker 3: for self government and learning. What you're pointing to is
• 6:29
  Speaker 3: also emphasized in the book, which is that algorithms can
• 6:34
  Speaker 3: echo chamber you. An algorithm might say, you know, you're
• 6:38
  Speaker 3: keenly interested in immigration, and you have this point of view,
• 6:42
  Speaker 3: so boy, are we going to funnel to you lots
• 6:44
  Speaker 3: of information because clicks are money, and you're going to
• 6:48
  Speaker 3: be clicking, clicking, clickling, click kicking, And that might be
• 6:50
  Speaker 3: a very good thing from the standpoint of the seller,
• 6:54
  Speaker 3: so to speak, or the user of the algorithm, but
• 6:56
  Speaker 3: from the standpoint of view, it's not so fantastic. And
• 7:00
  Speaker 3: from the standpoint of our society it's less than not
• 7:04
  Speaker 3: so fantastic, because people will be living in algorithm driven
• 7:09
  Speaker 3: universes that are very separate from one another, and they
• 7:13
  Speaker 3: can end up not liking each other very much.
• 7:16
  Speaker 2: Even worse than not liking each other. Their view of
• 7:19
  Speaker 2: the world aren't based on the same facts or the
• 7:23
  Speaker 2: same reality. Everybody knows about Facebook and to a lesser degree,
• 7:28
  Speaker 2: TikTok and Instagram and how it very much bulkanized people
• 7:33
  Speaker 2: into things. And we've seen that in the world of media.
• 7:37
  Speaker 2: You have Fox News over here, in MSNBC over there.
• 7:41
  Speaker 2: How significant of a threat does algorithmic news feeds present
• 7:48
  Speaker 2: to the country as a democracy, self regulating, self determined democracy.
• 7:56
  Speaker 3: Really significant? And there's algorithms and then there's large language models,
• 8:01
  Speaker 3: and they can both be used to create situations in which,
• 8:05
  Speaker 3: let's say, the people in some city let's call Los Angeles,
• 8:09
  Speaker 3: are seeing stuff that creates a reality that's very different
• 8:13
  Speaker 3: from the reality that people are saying, and let's say,
• 8:16
  Speaker 3: boise Idaho. And that can be a real problem for
• 8:20
  Speaker 3: understanding one another and also for mutual problem solving.
• 8:26
  Speaker 2: So let's apply this a little bit more to consumers
• 8:29
  Speaker 2: and markets. You describe two specific types of algorithmic discrimination.
• 8:36
  Speaker 2: One is price discrimination and the other is quality discrimination.
• 8:41
  Speaker 2: Why should we be aware of this distinction? Do they
• 8:44
  Speaker 2: both deserve regulatory attention?
• 8:47
  Speaker 3: So if there is price discrimination through algorithms in which
• 8:52
  Speaker 3: different people get different offers depending on what the algorithm
• 8:56
  Speaker 3: knows about their wealth and tastes, that's one thing, and
• 9:00
  Speaker 3: it might be okay. People don't stand up and cheer
• 9:03
  Speaker 3: and say hooray. But if people who have a lot
• 9:06
  Speaker 3: of resources are given an offer that's not as let's
• 9:09
  Speaker 3: say seductive, as one that is given to people who
• 9:14
  Speaker 3: don't have a lot of resources, just because the price
• 9:17
  Speaker 3: is higher for the roots than the poor, that's okay.
• 9:19
  Speaker 3: There's something efficient and market friendly about that. If it's
• 9:24
  Speaker 3: the case that people who are let's say, not caring
• 9:30
  Speaker 3: much about whether a tennis racket is going to break
• 9:33
  Speaker 3: after multiple uses, and other people who think that tennis
• 9:36
  Speaker 3: racket really has to be solid because I play every day,
• 9:40
  Speaker 3: and I'm going to play for the next five years.
• 9:43
  Speaker 3: Then some people are given the let's say, immortal tennis racket,
• 9:47
  Speaker 3: and other people are given the one that's more fragile.
• 9:51
  Speaker 3: That's also okay, so long as we're dealing with people
• 9:55
  Speaker 3: who have a level of sophistication. They know what they're
• 9:57
  Speaker 3: getting and they know what they need. It's the case
• 10:00
  Speaker 3: that for either pricing or for quality, the algorithm is
• 10:05
  Speaker 3: aware of the fact that certain consumers are particularly likely
• 10:10
  Speaker 3: not to have relevant information, then everything goes haywire. And
• 10:14
  Speaker 3: if this isn't frightening enough, note that algorithms are an
• 10:18
  Speaker 3: increasingly or an increasingly excellent position to know. This person
• 10:24
  Speaker 3: with whom I'm dealing doesn't know a lot about whether
• 10:28
  Speaker 3: products are going to last, and I can exploit that.
• 10:31
  Speaker 3: Or this person is very focused on today and tomorrow
• 10:35
  Speaker 3: and next year doesn't really matter. The person's present biased
• 10:38
  Speaker 3: and I can exploit that. And that's something that can
• 10:42
  Speaker 3: damage vulnerable consumers a lot, either with respect to quality
• 10:47
  Speaker 3: or with respect to pricing.
• 10:48
  Speaker 2: So let's flesh that out a little more. I'm very
• 10:52
  Speaker 2: much aware that when Facebook sells ads, because I've been
• 10:56
  Speaker 2: pitched these from Facebook. They could target an audience based
• 11:02
  Speaker 2: on not just their likes and dislikes, but their geography,
• 11:07
  Speaker 2: their search history, their credit score, their purchase history. Like,
• 11:11
  Speaker 2: they know more about you than you know about yourself.
• 11:15
  Speaker 2: It seems like we've created an opportunity for some potentially
• 11:21
  Speaker 2: abusive behavior. Where is the line crossed from Hey, we
• 11:27
  Speaker 2: know that you like dogs and so we're going to
• 11:30
  Speaker 2: market dog food to you to we know everything there
• 11:34
  Speaker 2: is about you, and we're going to exploit your behavior
• 11:37
  Speaker 2: biases and some of your emotional weaknesses.
• 11:42
  Speaker 3: Okay, So suppose there's a population of Facebook users who
• 11:46
  Speaker 3: are super well informed about food and really rational about food.
• 11:54
  Speaker 3: So they particularly happen to be fond of sushi, and
• 11:59
  Speaker 3: Facebook is going hard at them with respect to offers
• 12:03
  Speaker 3: for sushi and so forth. Now let's suppose there's another population,
• 12:09
  Speaker 3: which is they know what they like about food, but
• 12:12
  Speaker 3: they have kind of hopes and false beliefs both about
• 12:19
  Speaker 3: the effect of food on health. Then you can really
• 12:22
  Speaker 3: market to them things that will lead to poor choices.
• 12:25
  Speaker 3: And I've made a stark distinction between fully rational which
• 12:29
  Speaker 3: is kind of economic speak, and you know, imperfectly informed
• 12:34
  Speaker 3: and behaviorally biased people. Also economic speak, but it's really intuitive.
• 12:38
  Speaker 3: There's a radio show maybe This Will Bring It Home,
• 12:41
  Speaker 3: that I listen to when I drive into work, and
• 12:43
  Speaker 3: there's a lot of marketing about a product that is
• 12:48
  Speaker 3: supposed to relieve pain. And I don't want to criticize
• 12:52
  Speaker 3: any producer of any product, but I have reason to
• 12:55
  Speaker 3: believe that the relevant product doesn't help much. But the
• 13:00
  Speaker 3: the station that is marketing this product to people, this
• 13:04
  Speaker 3: pain relief product, must know that the audience is vulnerable
• 13:09
  Speaker 3: to it, and they must know exactly how to get
• 13:12
  Speaker 3: at them. And that's not in the interest of that's
• 13:15
  Speaker 3: not going to make America great again.
• 13:18
  Speaker 2: To say the very least. So we've been talking about algorithms,
• 13:23
  Speaker 2: but obviously the subtext is artificial intelligence, which seems to
• 13:29
  Speaker 2: be the natural extension and for the development of algos
• 13:34
  Speaker 2: tell us how as AI becomes more sophisticated and pervasive,
• 13:39
  Speaker 2: how is this going to impact our lives as employees,
• 13:43
  Speaker 2: as consumers, as citizens.
• 13:48
  Speaker 3: Chat GPT, chances are, knows a lot about everyone who
• 13:54
  Speaker 3: uses set. So I actually asked chat GPT recently, I
• 13:58
  Speaker 3: use it some not huge. I asked it to say
• 14:01
  Speaker 3: some things about myself, and it said a few things
• 14:05
  Speaker 3: that were kind of scarily precise about me based on
• 14:10
  Speaker 3: some number dozens, not one hundreds. I don't think of
• 14:13
  Speaker 3: engagements with chatch ept. So large language models that track
• 14:18
  Speaker 3: your prompts can know a lot about you, and if
• 14:22
  Speaker 3: they're able also to know your name, they can, you know, instantly,
• 14:26
  Speaker 3: basically learn a ton about you online, and we need
• 14:30
  Speaker 3: to have privacy protections that are working there. Still, it's
• 14:33
  Speaker 3: the case that AI broadly is able to use algorithms,
• 14:38
  Speaker 3: and generative AI can go well beyond the algorithms we've
• 14:42
  Speaker 3: gotten familiar with both to make the beauty of algorithmic engagement,
• 14:48
  Speaker 3: that is, here's what you like, here's what you want,
• 14:52
  Speaker 3: we're going to help you, and the ugliness of algorithms,
• 14:57
  Speaker 3: here's how we can exploit you to get you to buy.
• 15:00
  Speaker 3: And of course I'm thinking of investments too, so in
• 15:03
  Speaker 3: your neck of the woods, it would be a child's
• 15:05
  Speaker 3: play to get people super excited about investments, which AI
• 15:09
  Speaker 3: knows the people with whom it's engaging are particularly susceptible to,
• 15:14
  Speaker 3: even though they're really dumb, engagements.
• 15:17
  Speaker 2: Really really interesting. So since we're talking about investing, I
• 15:23
  Speaker 2: can't help but bring up both AI and algorithms trying
• 15:28
  Speaker 2: to increase so called market efficiency, and I always go
• 15:34
  Speaker 2: back to Uber's surge pricing. As soon as it starts
• 15:38
  Speaker 2: to rain, the prices go up in the city. It's
• 15:42
  Speaker 2: obviously not an emergency. It's just an annoyance. However, we
• 15:47
  Speaker 2: do see situations of price gouging after a storm, after
• 15:52
  Speaker 2: a hurricane, people only have so many batteries and so
• 15:54
  Speaker 2: much plywood, and they kind of crank up prices. How
• 15:58
  Speaker 2: do we determine what is the line between something like
• 16:02
  Speaker 2: surge pricing and something like, you know, abusive price gouging.
• 16:08
  Speaker 3: Okay, so you're in a terrific area of behavioral economics.
• 16:12
  Speaker 3: So we know that in circumstances in which let's say
• 16:16
  Speaker 3: demand goes up high because everyone needs a shovel and
• 16:22
  Speaker 3: as a snowstorm, people are really mad if the prices
• 16:26
  Speaker 3: go up, though it might be just a sensible market adjustment.
• 16:31
  Speaker 3: So as a first approximation, if there's a spectacular need
• 16:35
  Speaker 3: for something, let's say, shovels or umbrellas, the market inflation
• 16:41
  Speaker 3: of the cost, while it's morally abhorrent to many, and
• 16:44
  Speaker 3: maybe in principle morally abhorrent from the standpoint of standard economics,
• 16:49
  Speaker 3: it's okay. Now, if it's the case that people under
• 16:53
  Speaker 3: short term pressure from the fact that there's a lot
• 16:56
  Speaker 3: of rain are especially vulnerable and some kind of emotionally
• 17:02
  Speaker 3: intense state, so they'll pay kind of anything for an umbrella,
• 17:07
  Speaker 3: then there's a behavioral bias which is motivating people's willingness
• 17:11
  Speaker 3: to pay a lot more than the product is worth.
• 17:14
  Speaker 2: So let's talk a little bit about disclosures and the
• 17:18
  Speaker 2: sort of mandates that are required when we look across
• 17:21
  Speaker 2: the pond, when we look at Europe, they're much more
• 17:24
  Speaker 2: aggressive about protecting privacy and making sure big tech companies
• 17:29
  Speaker 2: are disclosing all the things they have to disclose. How
• 17:32
  Speaker 2: far behind is the US and that generally and are
• 17:36
  Speaker 2: we behind when it comes to disclosures about algorithms or
• 17:40
  Speaker 2: AI I.
• 17:42
  Speaker 3: Think we're behind them in the sense that we're less
• 17:45
  Speaker 3: privacy focused. But it's not clear that that's bad, And
• 17:50
  Speaker 3: even if it isn't good, it's not clear that it's terrible.
• 17:54
  Speaker 3: I think neither Europe nor the US has put their
• 17:58
  Speaker 3: regulatory finger on the actual problem. So let's take the
• 18:03
  Speaker 3: problem of algorithms not figuring out what people want, but
• 18:08
  Speaker 3: algorithms exploiting a lack of information or a behavioral bias
• 18:13
  Speaker 3: to get people to buy things at prices that aren't
• 18:16
  Speaker 3: good for them. That's a problem. It's in the same
• 18:20
  Speaker 3: universe as fraud and deception, and the question is what
• 18:23
  Speaker 3: are we going to do about it. A first line
• 18:25
  Speaker 3: of defense is to try to ensure consumer protection, not
• 18:29
  Speaker 3: through heavy handed regulation. I'm a long time University of
• 18:33
  Speaker 3: Chicago person. I have in my DNA not liking heavy
• 18:37
  Speaker 3: handed regulation, but through helping people to know what they're
• 18:42
  Speaker 3: buying and helping people not to suffer from a behavioral
• 18:47
  Speaker 3: bias such as, let's say, incomplete attention or unrealistic optimism
• 18:52
  Speaker 3: when they're buying things. So these are standard consumer protection
• 18:55
  Speaker 3: things which many of our agencies in the US, home
• 18:59
  Speaker 3: grown America. They've done that, and that's good and we
• 19:04
  Speaker 3: need more of that. So that's first line of defense.
• 19:07
  Speaker 3: Second line of defense isn't to say, you know, privacy, privacy, privacy,
• 19:12
  Speaker 3: though maybe that's a good song to sing. It's to
• 19:14
  Speaker 3: say right to algorithmic transparency. So this is something which
• 19:19
  Speaker 3: neither the US, nor Europe, nor Asia, nor South America
• 19:23
  Speaker 3: nor Africa has been very advanced on. So this is
• 19:27
  Speaker 3: a coming thing where we need to know what the
• 19:29
  Speaker 3: algorithms are doing. So it's public what's Amazon's algorithm doing?
• 19:35
  Speaker 3: That would be good to know, and it shouldn't be
• 19:38
  Speaker 3: the case that some efforts to ensure transparency invade Amazon's
• 19:43
  Speaker 3: legitimate rights.
• 19:45
  Speaker 2: Really, really fascinating, thanks Cas. Anybody who is participating in
• 19:51
  Speaker 2: the American economy and society, consumers, investors, even just regular
• 19:57
  Speaker 2: readers of news, needs to be away of how algorithms
• 20:01
  Speaker 2: are affecting what they see, the prices they pay, and
• 20:05
  Speaker 2: the sort of information they're getting. So with a little
• 20:09
  Speaker 2: bit of forethought and the book Algorithmic Harm, you can
• 20:13
  Speaker 2: protect yourself from the worst aspects of algorithms. And AI,
• 20:19
  Speaker 2: I'm Barry Redults you're listening to Bloomberg's at the Money