DeepMind Picks a Behaviouralist to Map the Post-AGI Economy

On 8 May 2026, Google DeepMind announced the hiring of Alex Imas as Director of AGI Economics, reporting into Shane Legg. Imas is a behavioural economist at Chicago Booth who recently co-authored a book with Richard Thaler updating the original Anomalies column. DeepMind has at least one parallel role open in London (Senior AI Economist), which suggests the team is being staffed out, not installed as a figurehead.

The hire deserves attention, and not just because of the title.

Behavioural, not macro

The choice of a behavioural economist (rather than a macro or labour economist) is itself a signal. Behavioural economics studies how humans actually decide under bounded rationality, not how idealised agents optimise on paper. Putting that lens on a post-AGI economy implies DeepMind expects the interesting questions to be about heterogeneous human responses, principal-agent problems with AI agents, and information asymmetries, rather than aggregate productivity curves.

Imas's recent working paper on "agentic economies" (visible on his personal site) already frames AI agents as inheriting and amplifying human heterogeneity. That is likely the seed of the research agenda, not a coincidence.

Competitive positioning

This is also a competitive move in a now-crowded space.

Anthropic publishes the Economic Index, which tracks real Claude usage across the labour market. OpenAI has economic research output of its own. DeepMind has now formalised an AGI Economics directorate explicitly oriented to post-AGI scenarios rather than near-term measurement.

The framing matters. Near-term impact research describes the present. Post-AGI economics tries to influence the policy and institutional response before the fact, while the assumptions are still up for grabs.

What to read instead of the announcement

If you want to follow this thread substantively rather than the press cycle, two sources are more useful than the news coverage.

The first is Imas's own Substack, which is where his actual thinking surfaces rather than being filtered through press releases.

The second is the Séb Krier essay Imas amplified in January 2026. Krier is AGI policy dev lead at DeepMind, and the essay argues against "full labour displacement" as the default assumption. Human labour share, in his view, will remain a substantial part of the economy a lot longer than AGI-maximalist timelines suggest, because complementarities persist and human involvement is often part of a service's value, not a deadweight cost.

That argument is probably close to the house view Imas will be building on. Worth reading before the next round of takes lands.

DeepMind Picks a Behaviouralist to Map the Post-AGI Economy

Behavioural, not macro

Competitive positioning

What to read instead of the announcement

Links