Is AI More Like a Mind or a Market

Bloomberg News recently published an article with that headline. Unfortunately that article is behind a paywall, but I found a gift link. I have previously posted about Alison Gopnik's (and others') view of large language models as cultural technologies that remix the products of human culture in productive ways. An excerpt:

Imagine a system that takes in billions of data points, conducts a massive number of simple but opaque operations, and then spits out responses that are both useful and, sometimes, destructive. The system seems all-knowing, but it is reductive and lacks common sense. We allow it to make major decisions, although a chorus of critics worries that it doesn’t share our values and may prove impossible to control. The system is not a large language model like ChatGPT. It’s the US stock market.

This is the sort of metaphor that a small but influential group of social and cognitive scientists say can help us better understand artificial intelligence. Today’s AI models are not, in their view, akin to a human mind. Rather, they’re a form of “cultural or social” technology that aggregates and passes on human knowledge — more like a printing press or even a bureaucracy or a market. If we want to understand how to manage AI, they say, we should study how we’ve handled new social technologies in the past.

Last year, Science published a version of this argument by Henry Farrell (a political scientist), Alison Gopnik (a psychologist), Cosma Shalizi (a statistician) and James Evans (a sociologist). “Beginning with language itself, human beings have had distinctive capacities to learn from the experiences of other humans and these capacities are arguably the secret of human evolutionary success,” the authors write. They go on to identify key ideas — from print to television to representative democracy — that transformed the nature of social learning by changing how societies process information.

The printing press, for example, didn’t just lower the price of making books — it overhauled how knowledge was built and allowed ideas to combine in new ways. “Once old texts came together within the same study, diverse systems of ideas and special disciplines could be combined,” the historian Elizabeth Eisenstein wrote in a 1968 paper. “Words drawn from one milieu and pictures from another were placed beside each other within the same books.”

The Science authors think we should view large language models along these lines — not as intelligence, but as a new form of cultural communication. They absorb everything ever written on the internet, among other media, and allow us to repackage and republish it in new ways. “Someone asking a bot for help writing a cover letter for a job application is really engaging in a technically mediated relationship with thousands of earlier job applicants and millions of other letter writers,” they write.

That’s not thinking, it’s remixing.