Global central bankers warn of dystopian AI future at Jackson Hole

The annual Jackson Hole Economic Policy Symposium, where the world’s most powerful monetary policymakers gather each August, delivered something genuinely unsettling this year. Princeton economist Markus Brunnermeier presented a paper arguing that artificial intelligence could eventually predict central bank moves better than the central bankers themselves, creating a financial system that’s harder to steer and more prone to chaos.

The paper, titled “Artificial Intelligence and the Brave New World in Finance,” was presented on August 27 and debated through August 29-30 at the symposium hosted by the Federal Reserve Bank of Kansas City.

The problem of asymmetric understanding

Brunnermeier coined a term that will likely haunt central banking circles for years: “asymmetric understanding.” The concept is straightforward. Imagine an AI system that can parse every Fed statement, every inflation data point, every labor market quirk, and synthesize a prediction about the next rate decision faster and more accurately than the humans making that decision have fully formed their own thinking.

That’s the scenario Brunnermeier laid out. AI systems that don’t just react to policy but anticipate it, front-running central bank decisions at machine speed.

Brunnermeier’s proposed solutions range from unconventional to borderline surreal. Central banks might need to hold separate press conferences for human audiences and AI audiences. They might need to become less transparent, not more, reversing decades of movement toward forward guidance and clear communication. In extreme scenarios, they might need to bypass markets entirely and intervene directly in credit allocation.

The optimist in the room

Not everyone at Jackson Hole shared Brunnermeier’s concern level. Fed Chair Kevin Warsh, delivering his keynote on August 28, took a decidedly more upbeat tone about AI’s economic trajectory. Warsh focused on the productivity gains that AI-related infrastructure investment could deliver, painting the technology as a potential catalyst for sustained economic growth.

The numbers behind the AI boom support Warsh’s enthusiasm, at least from a pure growth perspective. Leading AI labs are now generating over $100 billion annually in token sales, a figure that represents a more than 500% increase year over year.

Boston Fed President Susan Collins acknowledged the extreme scenarios Brunnermeier presented while emphasizing that policymakers need to take these risks seriously precisely because the technology is advancing so rapidly.

Why this matters beyond the mountains of Wyoming

Jackson Hole has historically been the venue where central banking orthodoxy gets challenged before it gets changed. Ben Bernanke’s 2010 speech there foreshadowed QE2. Mario Draghi’s 2014 appearance preceded the European Central Bank’s own bond-buying program.

Brunnermeier’s concerns aren’t entirely novel. Central bankers have grappled with the destabilizing effects of algorithmic trading for years, particularly after flash crashes demonstrated how automated systems can amplify market dislocations in seconds. But the current generation of AI represents a qualitative leap. Algorithmic trading follows rules. Modern AI systems learn, adapt, and optimize in ways their creators don’t always fully understand.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.