At the University we have weekly meetings for the Intelligent Systems Lab (ISL), that we cleverly call “LabMeets.” So, the topic today diverged from the usual talk on some aspect of computer intelligence, and was a brief talk followed by a short recording of an interview – the topic was essentially the Behavioural Economics of Daniel Kahneman. I was suitably impressed enough to write up some of my notes, and to exercise the liberal art of rhetoric. Although economics might at first seem quite distant from artificial intelligence, it is actually quite closely related.
As mentioned, the first section of today’s LabMeet was a brief overview of a paper by Daniel Kahneman and Amos Tversky entitled “Choices, Values and Frames” published in 1984 in American Psychologist. This paper showed that people tend to be “risk-averse” when the outcomes are seemingly positive, whereas “risk-seeking” when the outcomes are negative. A number of examples were given which highlight that given exactly the same scenario, the language of two different options has a direct relationship to the option chosen. This is a key aspect of decision-making, and therefore having an objective view allows for a more rational decision. It matters quite a lot to humanity, because it means that humans can very easily be manipulated, just through the use of language.
Kahneman also wrote Thinking, Fast and Slow. Which, as far as I am aware, goes into more detail about this particular theory. In the labmeet, we watched a youtube video of an interview with Kahneman regarding Thinking, Fast and Slow. Kahneman describes two systems which the brain uses:
- System 1: Is, in essence, the part of the mental processes which includes intuition and subconscious thought. It is that “gut feeling,” and is our fast response unit.
- System 2: Is, in essence, the more logical and rational mental processes. It (usually) takes more time to get a result from System 2 than it does from System 1, just because it takes time to calculate.
Considering one particular example. If somebody is on a short-term winning streak (e.g. in some kind of sport, or perhaps in playing the stock markets), our intuition (i.e. System 1) might tell use that that person is worth promoting or investing in. However, our rational mind (i.e. System 2) well tell us that statistically somebody that has been better than average for a lot longer (even if he/she is not currently having a winning streak), is better to promote or invest in. With this in mind System 2 is usually the better to go with.
However, we spend most of our lives, as humans, living in System 1 and it works for us most of the time. It is just that when the difficult decisions come, the result from System 2 will usually be the best decision. This is related to the regression towards the mean phenomenon.
What was also said was about leadership, and Kahneman mentions in the video that in terms of presidents of the USA; George Bush was more of a System 1 thinker, whereas Barack Obama is more of a System 2 thinker.
From my own point of view, I couldn’t help thinking that both System 1 and System 2 obviously have their pros and cons. I wonder, from a brain-improvement perspective, how this particular theory could help. Would it be possible to make our intuitions correct for more complex decisions (i.e. improving system 1 while retaining the speed), and would it be possible to improve our more rational brain by making our day-to-day lives more rational and making our rational thought incorporate the intuitive nature of our very being? It also leads me back to one of my previous blog posts where I consider how machines might deal with rationality and irrationality of humans.
Just some thoughts, and I’d be happy to hear yours…