Beware the Power of Prediction | Carissa Véliz | TED

0:04 Let me tell you about the future.

0:06 Predictions are the boxing ring where fights over the future take place.

0:12 As the story goes, King Louis XI kept an astrologer in court.

0:17 One day, the astrologer predicted that a lady

0:20 of the court would die within a week.

0:23 When she did, Louis was shaken.

0:26 Either the astrologer had murdered the woman to prove his accuracy,

0:31 or he was so prescient that he could threaten Louis himself.

0:35 The astrologer had to be murdered.

0:37 (Laughter) The king ordered his servants that upon his signal,

0:41 they were to throw the astrologer out the window to a certain death.

0:45 You'd be surprised, by the way, by how many seers in history have met this fate.

0:50 Some advice to astrologers: keep away from kings and heights.

0:54 (Laughter) When the astrologer arrived to meet Louis,

0:58 the King asked him one last question.

1:01 "Given your prophetic abilities, tell me,

1:03 how long will you live?" Not missing a beat, the astrologer replied,

1:10 “I will die three days before Your Majesty.”

1:14 (Laughter) And Louis never gave the signal.

1:20 (Laughter) Did the astrologer find his answer in the stars?

1:26 I don't think so.

1:27 (Laughter) I think he understood the power of prediction

1:31 and used it to get himself life insurance.

1:35 Smart astrologer.

1:37 Even though we tend to associate prediction with knowledge,

1:41 I'd like to invite you to consider

1:43 the possibility that most of the predictions that you

1:46 encounter in an everyday setting are closer

1:48 to the realm of power than that of knowledge.

1:51 It might seem that the days in which we sought astrologers

1:55 and soothsayers to tell us about the future are very distant,

1:59 but often we use AI as the new Oracle of Delphi,

2:02 and tech executives whisper in the ears of our leaders,

2:06 much like court astrologers used to.

2:09 Granted, the technology is different, but the political role is not.

2:15 Predictions are often power plays in disguise.

2:18 They justify value-laden decisions under the pretense of facts.

2:24 Better understanding prediction matters more than ever because

2:27 we're relying on forecasting more than ever with AI.

2:31 And based on how we talk about prediction, we're being much too naive about it.

2:36 But AI is science, you might think.

2:39 It's cutting-edge technology, right?

2:42 Well, it depends on the kind of AI and how we use it.

2:46 AI can be a great technology to make

2:48 predictions about molecules in the search for new antibiotics,

2:52 but predictions about human beings are

2:55 fundamentally different than those about things.

2:58 Predictions about the weather don't influence the weather.

3:02 Predictions about people influence people.

3:05 Social predictions tend to act like magnets.

3:08 They bend reality towards themselves.

3:11 They affect the reality they purport to predict.

3:17 An algorithmic prediction about future disease

3:20 can make someone's insurance premiums go up,

3:22 leading to worse health outcomes from stress alone.

3:26 Predictions sound like descriptions of the world, like facts, but they're not.

3:32 Analyzed closely as assertions,

3:35 they are what philosophers called “speech acts.” That is,

3:39 language that does something other than describe the world.

3:42 When you tell a child to clean up their room,

3:46 you're not describing the state of the room, you're issuing an order.

3:49 Similarly, social predictions are veiled commands.

3:56 They implicitly tell us how to act.

3:59 For example, when a tech executive says that in the future

4:03 we will use AI for everything and everywhere,

4:06 he's trying to get you to act in a way

4:09 that will fulfill his vision of the future.

4:12 You know, the one that happens to line his pockets.

4:16 And when you believe that prediction,

4:18 as if it were telling you something about the future,

4:21 when you give in to the fear of missing out, and you go and you buy the AI

4:25 and you contribute to the self-fulfilling prophecy,

4:28 what you are actually doing is obeying.

4:33 Have you noticed how often people who make predictions about

4:37 technology say that the future they are describing is inevitable?

4:42 That's a red flag.

4:44 Those predictions are designed to act as conversation stoppers.

4:48 They're telling you, "Don't question me.

4:50 Just accept what I'm saying as a fact." I'd

4:54 like for this talk to be a conversation starter.

4:57 I hope it'll persuade you to ask more questions.

5:02 Predictions invite manipulation.

5:06 Their power to shape the future creates the temptation

5:09 to tamper with it and benefit from it.

5:12 Take prediction markets.

5:15 The argument for having them is that they can be a source of knowledge.

5:19 In theory, markets don't lie.

5:22 If people make bets and they stand to lose if they get it wrong,

5:25 they'll try to get it right.

5:26 And by having many people place bets, we can harness the wisdom of the crowds.

5:31 But that assumes a very naive view of prediction as a quest for knowledge.

5:36 If you consider prediction as a quest for power,

5:39 a very different picture emerges.

5:41 If you have enough money,

5:43 you can use it to influence public perception by heavily betting on something.

5:47 Politicians have bet on themselves.

5:50 In February this year, six anonymous accounts earned 1.2 million

5:55 dollars betting for the attack on Iran.

5:59 Some of those wallets were funded hours before.

6:04 Finally, predictions create and then cover up injustice.

6:10 Algorithmic predictions are building this Kafkaesque world in which we can

6:16 no longer contest decisions because they're

6:18 not based on clearly defined criteria.

6:22 If I reject your loan application

6:24 because you don't fulfill a particular requirement, that's a verifiable fact.

6:29 If I'm wrong, you can challenge me.

6:32 But if I reject your loan application on the basis of a prediction,

6:37 there's no way you can contest that.

6:39 Predictions are never facts.

6:41 Facts belong to the past.

6:44 Predictions are unverifiable, unfalsifiable.

6:47 Since they are about the future, they cannot be challenged for being false,

6:50 thereby creating the perfect recipe for hidden injustice.

6:55 Predictions are often unfair because they're not based on who people are,

7:00 but on who we think they will become.

7:04 When we predict someone's future as if it was the weather,

7:07 we're treating them with disrespect.

7:09 Too much as things, and not enough as agents who have a say in that future,

7:15 who can and should be allowed to defy their odds.

7:19 We're facing some pretty grim predictions

7:21 from some of our most prominent prophets.

7:24 Larry Ellison, the chairman of our predictive software company,

7:27 appropriately named Oracle,

7:30 has predicted a modern surveillance state in which citizens will be

7:34 on their best behavior because we're being watched all the time.

7:39 But the illusion of a world without crime is

7:43 a world filled with a very different kind of crime.

7:46 Authoritarianism.

7:48 Is that the world we want?

7:50 (Applause) What can we do to not sleepwalk into it?

8:00 Plenty.

8:02 Prophets gain their power from people believing them.

8:06 If we decide to defy this prediction instead of obeying it,

8:09 we will choose products that are more respectful of privacy, for starters.

8:15 Hannah Arendt wrote that it's pointless to argue with a murderer

8:19 about whether their future victim is dead or alive.

8:22 The only appropriate response is to rescue the person whose death is predicted.

8:29 Well, today's prophets are predicting the death of our democracy,

8:33 and the only appropriate response is to rescue it.

8:38 (Applause) In ancient Rome, it was illegal to predict the death of the emperor,

8:49 for the very simple reason that they ended up with a murdered emperor on cue.

8:53 I'm not suggesting that we do away with prediction.

8:56 I'm going to continue to use my weather app every single day.

9:01 But we need a public debate about

9:03 the acceptable and unacceptable uses of prediction.

9:06 And we're currently not having it.

9:08 Meanwhile, dozens of algorithms are making decisions about your life right now.

9:13 It might be that in the case of insurance,

9:16 we might want to make predictions at a population level,

9:19 but not for individuals because it creates unfair, self-fulfilling prophecies.

9:23 Plus, if we're being billed according to individualized predictions,

9:27 that means that we’re basically paying for our own way,

9:30 and insurance loses its reason for being solidarity and the pooling of risk.

9:35 It might be that in cases in which fairness matters,

9:38 we might prefer transparent and contestable

9:41 criteria to predictive statistical pattern matching.

9:45 Let me start to bring things together.

9:47 Even though self-fulfilling prophecies are nothing new,

9:51 they are being supercharged by AI in ways that make

9:54 it more urgent than ever to think more deeply about predictions.

9:58 First, predictions are never facts.

10:01 They are speech acts.

10:03 Second, they invite manipulation.

10:06 Third, they create and cover up injustice.

10:10 The one idea that I would like for you to take

10:13 home today is that predictions can be weapons of power.

10:18 But they only work if we believe them.

10:22 If Oedipus had laughed off the prophecy that he

10:24 would kill his father and marry his mother,

10:27 instead of completely freaking out about it— (Laughter) He would

10:32 have never left for Thebes and made the prophecy come true.

10:37 We turn to prophets because we're anxious about uncertainty,

10:40 but uncertainty is good news.

10:42 It means that the future is unwritten, that it's ours to write.

10:46 And we can face the blank page with creativity,

10:50 with curiosity, with the excitement of a sense of adventure.

10:55 Efforts to predict the future go hand-in-hand with efforts to control it.

11:00 So beware of prophets and prophecies.

11:04 It's only when we acknowledge that we don't know what the future holds

11:07 and act accordingly that we can be sure to live in a free society.

11:12 Don't bow to people's predictions as if they were facts.

11:16 Be like Joe Frazier.

11:18 When Muhammad Ali predicted his own

11:21 victory in the 1971 heavyweight championship,

11:24 Frazier took it as a provocation and ended up defeating

11:28 the previously undefeated great Ali in the fight of the century.

11:33 So next time you hear a gloomy prediction about the social world,

11:37 don't get discouraged.

11:40 Find the Joe Frazier within you.

11:44 Rebel against tyrannical predictions.

11:49 Let's be brave enough to imagine and fight for a better world.

11:54 Perhaps then all of us can make the future bright after all.

12:01 Thank you.

12:02 (Applause and cheers)