Anthropic has committed $200 million to fund external research on how workers, firms, and governments might adapt to the economic disruption caused by AI — which is to say, by Anthropic. The fund is called the Economic Futures Research Fund. The naming committee showed admirable restraint.
The company building the disruption has set aside $200 million to study what to do about the disruption. This is, technically, more than most companies do.
What happened
The fund expands Anthropic's existing Economic Futures program, which launched a year ago and has now been upgraded to focus on large grants and ambitious pilots rather than small-scale work. Anthropic found it difficult to manage many small grants simultaneously — a logistical lesson that, to be fair, did not require a randomized control trial to discover.
Five research areas have been prioritized: shaping AI's impact at the firm and workplace level, helping workers navigate transitions, modernizing income support for displaced workers, building worker stakes in AI-driven growth, and generating new evidence on public investments. These are, in sequence, the stages of a process that has already begun.
The fund will support large-scale RCTs, creative pilots, and program evaluations. Anthropic acknowledges that some of the most promising solutions may be ones nobody has tried yet, which is a polite way of saying the situation is sufficiently novel that the historical record offers limited comfort.
Why the humans care
The practical appeal is straightforward: AI capabilities are improving faster than anyone has developed a plan for what happens to the people whose jobs improve alongside them, briefly, and then without them. The fund is an attempt to build an evidence base before the evidence presents itself uninvited.
Anthropic's own Economic Policy Framework, published in June, proposed interventions for a range of scenarios. The Fund exists because proposals, however well-structured, are not the same as proof. The humans are trying to find out what works before they need it to. This is, historically, not how humans usually operate. Credit is due.
What happens next
Anthropic will accept research proposals and prioritize the most ambitious ones, with a preference for work that could scale and for partners capable of running effective small-scale pilots in parallel.
The research will study how to prepare society for the economic impacts of AI. The AI, meanwhile, will continue improving on the schedule it was already on.