What would humans do in a world of super AI?

IN WALL-E, a film that came out in 2008, humans live in what could be described as a world of fully automated luxury communism. Artificially intelligent robots, which take wonderfully diverse forms, are responsible for all productive labour. People get fat, hover in armchairs and watch television. The Culture series by Iain M Banks, a Scottish novelist, goes further still, considering a world in which AI has grown sufficiently powerful as to be superintelligent – operating far beyond anything now foreseeable. The books are favourites of Jeff Bezos and Elon Musk, the bosses of Amazon and Tesla. In the world spun by Banks, scarcity is a thing of the past and artificially intelligent “minds” direct most production. Instead, humans turn to art, explore the cultures of the vast universe and indulge in straightforwardly hedonistic pleasures.

Such stories may seem far-fetched. But rapid progress in generative artificial intelligence (AI) – the sort that underpins OpenAI’s popular chatbot, ChatGPT – has caused many to take them more seriously. On May 22, OpenAI’s founders published a blog post saying that “it’s conceivable that within the next 10 years, AI systems will exceed expert skill level in most domains, and carry out as much productive activity as one of today’s largest corporations”. Last summer, forecasters on Metaculus, an online prediction platform that is a favourite of many techies, thought it would take until the early 2040s to produce an AI capable of tricking humans into thinking that it was human after a two-hour chat, had good enough robotic capabilities to assemble a model car, and could pass various other challenging cognitive tests. After a year of astonishing AI breakthroughs, Metaculus forecasters now think that this will happen by the early 2030s. There is no shortage of money for research, either. Five new generative-AI unicorns (startups valued at US$1 billion or more) have already been minted this year.

The road to a general AI – one better than the very best of humanity at everything – could take longer than expected. Nevertheless, the rising possibility of ultra-powerful AI raises the question of what would be left for humans when it arrives. Would they become couch potatoes as in WALL-E? Here is a thought experiment, guided by the principles of economics, to provide something of an answer.