In a recent blog post, Basant Dar There were some interesting comments made about the potential role of AI in Fed policy.
Can AI help us better understand data and design better economic policy interventions?
Last week, I had lunch with several central bankers who wanted to talk to me about AI. They were interested in discussing how to think about the risks and opportunities associated with the growing presence of AI in financial markets. I asked them the same question as Paul. Can AI do the job of a central bank? During the meal, they realized that AI certainly has more to offer than humans when it comes to connecting the dots. Dot contains historical data on all past central bank monetary policy decisions, past communications and forecasts, all historical and contemporary data available at the time, and all literature published on the subject over time. It is included. Basically, we felt that AI might be better at modeling the economy. At a minimum, the model should be compared to the model currently in use.
But humans still need to take responsibility, they concluded. After all, you can’t get mad at the AI if it’s wrong.
I can certainly see the appeal of using AI to formulate plans. Financial policy. But overall, I still prefer to rely on a market approach to policy.
Can AI make better predictions than the NGDP futures market? That’s hard to say. AI predictions can incorporate not only market predictions but also other factors missed by market participants. In that case, you can think of AI prediction as a human-machine hybrid. In the early days of computer chess, plays that combined the insights of both grandmasters and computer programs were superior to humans or machines working alone.
However, if AI is truly better than markets in certain respects, we would expect market participants to use AI when engaging in trading in financial assets. Various AI insights will then be incorporated into market prices.
From this perspective, the addition of AI is not a qualitative change in markets or forecasts. Rather, the AI revolution will add a large number of really smart “entities” to the market, making it even more efficient than before. While this may be important, it is still a quantitative rather than qualitative change. Even if some AI is smarter than any individual, including Jay Powell, no AI is smarter than a market that includes many smarter humans and many smarter AIs.of wisdom of the crowd Still holding it.
A potentially more promising use of AI is coming up with appropriate targets for Fed policy (e.g., inflation versus NGDP growth, or levels versus growth rates). These questions cannot be answered by market forecasting. Future data will clearly resolve the question of who was right and who was wrong. Of course, this applies to a wide range of public policy issues. So you can imagine asking an AI whether the death penalty increases or decreases society’s total utility. If thinking about this makes you feel nauseous, you may find some comfort in thinking about the following: AI is currently far from being able to answer those kinds of questionsand it’s not clear whether they can do so.
P.S. david beckworth We asked AI what he thought about the use of AI in monetary policy.
(Number of comments: 5)
