:quality(75):focal(1556x300:1566x310)/cloudfront-us-east-1.images.arcpublishing.com/bloomberglinea/LRSZ4IVHZJBQ3AKLY7I6UBENWI.jpg "Daniel Kahneman is considered one of the fathers of behavioral economics and won the Nobel Prize in Economics in 2002 \"for integrating insights from psychology research into economic science, especially as it relates to human judgment and decision making under uncertainty.\"")
Bloomberg Línea — CEOs and senior executives do not occupy such positions for nothing: they accumulate many years of experience, play a leadership role, and have technical and social-emotional skills. But in the near future, they will have to live with a reality in which strategic business decisions within corporations will also be made by artificial intelligence (AI), according to Daniel Kahneman, winner of the Nobel Prize in economic science in 2002, and considered one of the fathers of so-called behavioral economics.
“It won’t be long before artificial intelligence is better than people because it learns faster (...) So we can expect that there will be more and more areas where artificial intelligence will become more and more important,” Kahneman said in an exclusive interview with Bloomberg Línea by videoconference from the United States.
:quality(75)/cloudfront-us-east-1.images.arcpublishing.com/bloomberglinea/YK3FQH6JOFHPPLDJO3T3RKDYQY.jpg "Glass Half-Full: VC Funds Keep Raising From LPs to Invest in LatAm")
“It will be possible to develop artificial intelligence that can evaluate business proposals at least as well or possibly better than a CEO. There will be a lot of decisions made by artificial intelligence. It hasn’t happened yet, but I think that moment is coming,” said the emeritus professor of psychology and public relations at Princeton University, who also predicted that there will be a lot of resistance from business leaders who today make the major decisions.
Kahneman, 88, will travel to Brazil at the end of the month to participate as the main speaker at Data Driven Business, a data analytics event promoted by Neoway in partnership with B3 (B3SA3). The event, to be held August 30 in São Paulo, will have as its theme the use of intelligence in data analysis for companies’ decision-making processes.
Despite the increase in knowledge on the subject and the advancement of technology over the years, companies still don’t have, in general, decision-making processes designed for this purpose, based on a series of processes, Kahneman said.
“We like people who make decisions quickly. We like people who are very decisive and who decide on instinct, and that’s quite popular with political leaders, certainly, but also with business leaders,” Kahneman said, and who won the Nobel Prize “for integrating insights from research in psychology into economic science, especially with respect to human judgment and decision-making under uncertainty,” according to the Swedish Academy’s statement.
:quality(75)/cloudfront-us-east-1.images.arcpublishing.com/bloomberglinea/QDWD3GBAEJAEFEH7YO3NG25BDM.jpg "L’Attitude Ventures Wields $100M to Invest in Latino Entrepreneurs in the US")
Kahneman published a new book, “Noise - a Flaw in Human Judgement”, together with Olivier Sibony and Cass R. Sustein in 2021, and which deals with biases that interfere with decision making and lead to a variability of outcomes - such as different sentences decided by more or less severe judges facing the same case in a court of law.
This is what he and the other authors describe as ‘noise’ in different environments. It is a failure that can cost millions of dollars to companies and society due to the inefficiencies generated, such as not assertively pricing the premiums of an insurance company.
The following inteview was edited for length and clarity:
Bloomberg Línea: Your research and books have attracted attention for decades. Are we, as a society, making better decisions than we were 20 or 30 years ago?
Daniel Kahneman: I really don’t know enough to give a good answer to that question. My guess is that organizations have improved, especially in terms of the information they use.
Are we therefore more prepared? There is more knowledge about the process and, more recently, we can rely on technology and the use of intelligence in data analysis.
I think so, certainly. In the book we published, Noise, we propose ideas to improve decision making in organizations. These are ideas that seem feasible, not impossible. I think people are concerned about the quality of decision making and are trying to make it a good thing.
Are there sectors or organizations, private or public, that stand out for the way they are making decisions?
Well, I’m an academic, so I live in an ‘ivory tower’ and haven’t had contact with many organizations. But clearly, first of all, it depends on the areas within an organization. There may be one area that is effectively good, and others where the decision making is not so good.
:quality(75)/cloudfront-us-east-1.images.arcpublishing.com/bloomberglinea/JEKNINKCFNEOHK2WY6SVIFHE3I.jpg "VCs Advise Startups Running Out of Cash: Try in the M&A Markets")
An example of very good organizational decision making is Google’s hiring process, which is very impressive, from what I’ve heard. And I have seen very good decision-making in venture capital firms, but the key here is whether the decision-making process is totally accidental and developed over time or whether it is designed in the best possible way to optimize results.
In general, I don’t think the decision-making of organizations is something that is very designed, mainly because historically they have acted one way and coached one issue or another, and that has evolved over time. But decision making is not designed based on knowledge of how to organize meetings or how information is passed from one person to another. Design is still not very prevalent.
In your evaluation, what is the main obstacle for companies, organizations, and even people to make better decisions?
Whenever there is a change, whenever you want to reform a process, there will be some people who will lose out because they were better off with the previous system. And the problem, and this is really a very common problem, is that the people who are going to lose from the change usually resist and struggle more than the people who are going to gain something. We call this loss aversion.
And this hinders any change. Loss aversion is a general hindrance to any effort aimed at any change. And it is generally true that potential losers fight hard not to lose, and potential winners fight a little less. And the interests of institutions are sometimes absurd.
And there is also a widespread preference for intuitive decision-making over reasoning. We like people who make decisions quickly. We like people who are very decisive and who decide on instinct, and that’s quite popular with political leaders, certainly, but also with business leaders.
Great business leaders have their intuition and flashes, and that tradition of how to get things done. So there is a real conflict between serving intuitive thinking and making decisions quickly and the more systematic or slower approach.
And the slower approach has a big problem. It tends to be bureaucratic. Because there are rules and procedures, there is resistance. There is people’s preference for intuition and resistance to any formalized procedure. No criticism of the preference for intuition, but I think that’s really, in a way, unfortunate.
There is a real risk that when we adopt procedures that aim to increase reliability in decision making, it ends up creating bureaucracy, an administrative or mechanical mentality.
So whenever there is an attempt to improve the quality of decision making, you have to keep this in mind: the process, whatever it is, has to feel natural, so that people feel that it will help them. People have to feel that they are participating in this process and that it has helped them achieve something they want, that they have performed better.
Are there areas that are supposed to have better decision making processes? Like the judiciary and medicine?
Obviously, in the justice system and in the medical system, there is a lot of noise, and reducing it would be very important if you can overcome the resistance. There is a lot of resistance to reducing the noise. In general, it would be easier to adopt better procedures in situations where decisions are made simultaneously.
In venture capital, when the investor is possibly evaluating startups, for example, there should be some questions that they always ask in a particular way and some mistakes from which to try to move forward. So that’s one direction to go. Decisions that repeat themselves represent a good way to start improving. Ideally, you would want to have an improved process for the most important decisions, but that would be very difficult.
The processes would start at the top of a company, with the boards of directors and top executives, and go down the hierarchy from there. But most likely the procedures would be improved not at the top level, but at the somewhat lower level. And this is because the people at the top really want to remain free to make their own decisions, because they will have their own intuitions and will not be bound by procedures.
:quality(75)/cloudfront-us-east-1.images.arcpublishing.com/bloomberglinea/M22FQQ3PJ5C5LNS66UEWFLZPXY.jpg "Spain Looks to Attract Latin American Workers As its Population Ages")
This is an interesting point. In the pandemic, authorities had to make decisions about lockdowns and mandatory vaccination. CEOs had to decide on the working hours of their employees. How do you evaluate the processes that were adopted to make these decisions?
Well, it is very clear that when many people are faced with the same situations and make very different decisions, something is probably not perfect. This is really a matter of evaluating teamwork. It’s a matter of evaluating the information that was available and how they used it. And some differences may be related to differences in the population that these people were dealing with. But certainly a lot of variability can simply indicate noise in the decisions.
In your latest book, you point out the difference between ‘noise’ and ‘bias’. Could you explain?
Bias is a mistake that many people make that creates a general bias. And there are some biases that we know are against groups. That’s what people usually think of when they talk about bias. But there are biases that are completely different. For example, there is an optimistic bias, of people who are quite optimistic in evaluating issues that they are involved with. That is a bias. There are many biases that are familiar.
Noise is something completely different. Noise is variability. It is something that people don’t share. So bias is something they have in common, while noise is something they don’t have in common. And in general, the best way to understand noise is to think that different people have different biases. And it is the variability in people’s biases that produces noise.
So when there are some optimists and some pessimists, or when there are some strict judges and some less strict judges, that produces noise in the system. You can imagine, for example, the judiciary: the whole system being very punitive, with very harsh sentences, maybe harsher than necessary. That could be called bias, but when different judges vary in their sentencing and express their own individual biases in that way, that is noise.
And how do we mitigate the effects caused by ‘noise’?
Well, what people want in general, and this is the problem I mentioned earlier, is to avoid procedures and bureaucracy. People want their thinking to be organized without being completely limited and restricted. You want some discipline in thinking. And discipline, if it is adopted by many people, will reduce noise. So there are ways of dealing with problems, like breaking them down into dimensions, which will probably reduce the noise.
We know that thinking comparatively is a way to reduce noise in relation to making judgments as if they were single cases. Comparing cases is a better thing to do. People do better at comparisons and agree more with each other and with themselves. So those are some of the procedures that people can look at.
:quality(75)/cloudfront-us-east-1.images.arcpublishing.com/bloomberglinea/ZZ2OS5F5LNEHJEYFNIPSR47SHY.jpg "Which Latin American Country Ranks Top for Remote Working?")
Is there a lot of noise in the financial markets as well?
Well, that is a very different kind of system. The financial market works on differences of opinion. So it is not like an insurance company or a medical system where you want doctors to agree on a diagnosis. In the market, you don’t necessarily want people to be inside financial companies. There is a lot of knowledge. So different analysts don’t necessarily come to the same conclusions, and they don’t necessarily agree in their assessments of the prospects.
Can these different points of view be balanced in a way that improves the decision-making process?
If you take this more systematic approach, for example when dealing with a problem by breaking it down into dimensions, this is a process that is likely to reduce noise. What you are really trying to do is to defer intuition. That is, you want people to think first and accumulate a lot of information, and then they can rely on intuition, but you try to postpone it.
As you can see, I have a bias against intuitive decision making. It is inevitable and ultimately you should trust intuition, but not too quickly, and don’t trust intuitions too much. So there’s a lot of that and it’s good to try to balance it.
Can technology help us achieve that balance point?
Well, technology will help a lot. First of all, data is organized. And artificial intelligence is getting into more and more areas of decision-making, actually making decisions or proposing them. And in the areas where that happens, when it becomes possible, artificial intelligence comes close in terms of performance to humans.
It won’t be long before artificial intelligence is better than people. because it learns faster. Artificial intelligence accumulates information much faster than any individual, and this is a great advantage. So we can expect that there will be more and more areas where artificial intelligence will become more and more important.
Will we be prepared for this transition and to accept this help from artificial intelligence?
I think it will be very difficult and there will be a lot of resistance. And there will be resistance at the point where very important decisions that are the responsibility of important people can be made by artificial intelligence. One example is the evaluation of business proposals that are complicated. This is something that we generally rely on very experienced executives for, because they are the leaders, they are very well paid, and they owe their position to that. People expect them to make good decisions.
And that may happen in a decade or two: it will be possible to get artificial intelligence that can evaluate business proposals at least as well or possibly better than a CEO. There will be a lot of decisions made by artificial intelligence. It hasn’t happened yet, but I think that moment is coming.
It is very clear that you still maintain interest and curiosity about different topics. What has attracted your attention the most?
Well, you can see that I am very interested in artificial intelligence. And the progress in that domain. I follow it more closely. If I could, I would learn a lot about the brain, because that is another topic that is extremely important, but I am very behind in understanding what is going on. But what fascinates me most at the moment is probably AI.
:quality(75)/cloudfront-us-east-1.images.arcpublishing.com/bloomberglinea/BPBFBUJVUJDATAFGS3G6J6EKLM.jpg "Global and Local Companies in Corporate Investment Pullback")
In Michael Lewis’ book The Undoing Project one of the main messages is about how powerful and inspiring collaborative work can be, from his decades-long story with Amos Tversky. We are currently seeing the rise of remote study and remote work. Do you have an opinion about the changes in the way we work?
Well, my guess is that there are certain things that can only be done much better in person than remotely. And this is especially true, for example, for creating friendships. Doing it remotely is not exactly the same as face to face. But once the friendship is established, you can collaborate remotely.
In writing Noise, I worked in collaboration with Olivier Sibony. Before Covid, we were traveling to Paris and New York to work together. And then it became impossible because he was in Paris, and I was in New York. And we started using Zoom for an hour every day and it was much more productive. So that question is going to vary a lot. My guess is that creating new collaborations and partnerships will be more difficult remotely. But for many other things, remote work will be more efficient.
Translated from the Portuguese by Adam Critchley
