Last week, the managing director of the IMF, Kristalina Georgieva, wrote a blog post in which she described how “expectations may run ahead of productivity gains” and how, in “a severe scenario, we could see a market correction that cascades through the highly leveraged and circular AI ecosystem” as well as drawing attention to the potential for ‘profound’ job market disruption and how AI cybersecurity threats are becoming a source of systemic risk.
This echoes earlier warnings from Georgieva that, on the one hand, models such as Anthropic’s Mythos “can be used to destroy the financial system” if bad actors get their hands on them and, on the other, an AI bubble bursting would be disastrous for the financial system.
There are many other recent warnings. Katie Martin, the FT’s Markets columnist, has an excellent series of articles in recent weeks describing such concerns. Also, this past week, on Ezra Klein’s excellent podcast, as part of a debate, Paul Kedrosky highlights the many reasons to be concerned about a bust, likening AI to other past general-purpose technologies such as the railways or the canals or electrification that also vastly overbuilt, leading to bankruptcies of some of the wealthiest firms and individuals.
Are you convinced by these warnings? If so, what worries you most? If not, why not?