When one of the most cited economists in growth theory tells you there’s a one-in-three chance AI causes human extinction, it’s worth paying attention. It’s worth paying even more attention when the company building that AI hires him to study the problem from the inside.

Chad Jones, the STANCO 25 Professor of Economics at Stanford Graduate School of Business, is taking a leave of absence to join the newly established Anthropic Institute. His start date is June 30, 2026, and his mandate is straightforward if somewhat existential: figure out what AI actually means for the future of the economy, and whether that future includes humans at all.

The math behind the existential bet

Jones’s recent NBER working paper, titled “A.I. and Our Economic Future,” lays out the intellectual framework that likely caught Anthropic’s eye. The paper models AI-driven acceleration in research and innovation, using log utility assumptions to calculate the trade-offs humanity faces.

The numbers are striking. Jones’s model produces a one-in-three probability of human extinction from superintelligent AI systems. The flip side: a two-thirds chance that living standards increase by roughly 55 times current levels.

The paper doesn’t argue for stopping AI development. Instead, it highlights that faster research growth, while enormously promising, mechanically increases the probability of catastrophic misuse. Speed and danger are, in Jones’s framework, structurally linked.

Why Anthropic built an institute

The Anthropic Institute launched on March 11, 2026, led by company co-founder Jack Clark. Its stated goal is to investigate the systemic effects of AI on society, the economy, and governance.

Jones isn’t the first economist to work at Anthropic. The company has previously brought on economists like Anton Korinek, signaling that its leadership views economic analysis as central to understanding AI risk, not peripheral to it.

What this means for investors

Jones’s presence at Anthropic could influence policy discussions in Washington and Brussels. Academic economists with his stature tend to get invited to testify before congressional committees and advise regulatory bodies. His models quantifying extinction risk alongside economic upside give policymakers something they rarely have in technology debates: a formal framework for thinking about trade-offs, rather than vibes and anecdotes.