Richard Thaler, Behavioral Economics and A.I.

Richard Thaler is the third Behavioral Economics giant to be awarded a Nobel Prize in Economics after Daniel Kahneman (2002) and Robert Schiller (2013). I found his intellectual autobiography very enlightening:

 Here's the 10/9/2017 Nobel Prize press release, from which I quote this fascinating tidbit:

Lack of self-control: Thaler has also shed new light on the old observation that New Year's resolutions can be hard to keep. He showed how to analyse self-control problems using a planner-doer model, which is similar to the frameworks psychologists and neuroscientists now use to describe the internal tension between long-term planning and short-term doing. Succumbing to shortterm temptation is an important reason why our plans to save for old age, or make healthier lifestyle choices, often fail. In his applied work, Thaler demonstrated how nudging – a term he coined – may help people exercise better self-control when saving for a pension, as well in other contexts.

and here's a quote from a New York Times article:

Professor Thaler has played a central role in pushing economists away from that assumption. He did not simply argue that humans are irrational, which has always been obvious but is not particularly helpful. Rather, he showed that people depart from rationality in consistent ways, so their behavior can still be anticipated and modeled.

Richard Thaler is one of the fathers of Behavioral Economics, whose basic assumption is that human beings, far from being the Homo Economicus of orthodox, neo-classical economics, can behave in irrational ways. Tangentially related is the subject of Artificial Intelligence. A commentator in the New York Times article mentioned above, Bey Melamed posted: "I wonder how Thaler's theories clash with / are supported by or commensurate with artificial intelligence. If / when AI has reached sufficient sophistication it should be able to (or even measured by the ability to) make human decisions with appropriate irrationality sprinkled in, no?"

My response to this comment is that alternatively, in a more sinister vision of the future, AI could be programmed to follow all the rules and behaviors of Homo Economicus, never showing the irrational idiosyncrasies exhibited by humans. Since we have a huge theoretical framework for just such creatures, it would make sense to replace human agents by algorithms across many domains. It is already the case in Financial Markets, where algorithmic, non-human traders have been taking over many tasks human traders used to perform. That would, of course, accelerate the replacement of humans by programs and algorithms, a pervasive process already under way. A failure to adapt to this trend will result in negative outcomes such as mass unemployment, community disintegration and social upheaval.

Here's another great New York Times article (from a 2001 NYT magazine profile) on Richard Thaler, written by Roger Lowenstein, the author of When Genius Failed, the definitive book on the saga of the hedge fund Long Term Capital Management (LTCM, 1994-1998):

Thaler spearheaded a simple but devastating dissent. Rejecting the narrow, mechanical homo economicus that serves as a basis for neoclassical theory, Thaler proposed that most people actually behave like . . . people! They are prone to error, irrationality and emotion, and they act in ways not always consistent with maximizing their own financial well being.