The most interesting question about AI isn’t which model just topped a benchmark. It’s this: what actually happens to us — our jobs, our incomes, our society — when these systems really land everywhere? On July 22, Anthropic gave an answer you can measure in dollars: $200 million goes into the Economic Futures Research Fund, and to kick it off, there’s a detailed research agenda.
What it’s about
The fund backs external research — universities, research institutes, non-profits — into which interventions make an economy more flexible and resilient when AI reshapes it. Not small studies, but big swings: Anthropic wants to fund projects mainly in the $5 to $30 million range, and deliberately won’t finance anything under a million. The reasoning is refreshingly honest — the predecessor program struggled to manage lots of small grants, and the truly useful insights come from large-scale field experiments and pilots.
The five priorities
Anthropic is setting priorities. First: how does AI change work at the firm level? Field experiments should reveal which collaboration patterns preserve human expertise — and who ends up capturing the productivity gains. Second: helping people through transitions — retraining, job placement, new apprenticeship models for when entry-level tasks suddenly get done by AI. Third: modernizing income support — our system assumes joblessness is temporary. What if it isn’t anymore?
Fourth, and this is the most intriguing one: giving workers a stake in the gains before the disruption arrives. Anthropic is thinking out loud about capital accounts for everyone, equity-sharing, AI-sector dividends. Fifth: public investment — pilots that fund human-facing work, from libraries to tutoring to community health.
My take
You could read this cynically — the company helping drive the disruption is funding the research on how to cushion it. The thought is obvious enough. But I still think it’s the right move. Because the uncomfortable truth is this: we have almost no solid data on what actually works in an AI-transformed economy. We have opinions, ideologies, gut feelings — but few real field experiments.
What I like is the openness to things nobody has tried yet. Anthropic writes that we might be entering a moment without historical precedent, where the most promising solutions are the ones no one has tested. That’s a rare dose of humility. Whether it turns into useful policy won’t be decided on a blog page — it’ll be decided over the coming years, in real pilots with real people. But the start has been made, and it’s a big one.
Sources: