Anthropic has released two policy frameworks aimed at preparing the United States for the economic and governance challenges it says could come with rapidly advancing artificial intelligence. The company paired the proposals with $350 million in new funding commitments for research and workforce-related programs.
The first framework focuses on what Anthropic calls advanced AI governance. It calls for stronger transparency requirements, independent evaluation of capable systems, and government authority to stop or slow deployments that could pose serious risks. The second framework addresses the economic side of AI, arguing that policymakers should prepare for possible labor market disruption and design ways to spread the gains from AI more broadly.
Anthropic said the two documents are meant to address related questions. One asks how to steer increasingly powerful systems safely. The other asks how workers and institutions should be protected if AI reduces demand for human labor.
In its economic framework, Anthropic says it cannot forecast exactly how quickly AI will affect jobs, but believes policymakers should prepare for several possible outcomes. The company sets out recommendations for three broad scenarios: about 5% unemployment, 10% unemployment, and a far more severe case of unprecedented unemployment.
The company says it is not seeking job displacement and wants to minimize it. Still, it argues that some displacement may be unavoidable and that governments should be ready to respond. Among the ideas it highlights are better labor-market data collection, stronger analytical capacity inside government, and modernization of benefit systems such as unemployment insurance so they can scale faster if needed.
Anthropic also says the response should vary depending on the severity of disruption. In a moderate disruption scenario, it suggests measures such as workforce training grants, licensing reform, wage insurance, and incentives for employers to retrain or redeploy workers rather than cut staff. If unemployment rises more sharply, the company calls for expanded unemployment support, sector-specific transition aid, and basic-needs relief.
In the most extreme scenario, Anthropic says policymakers may need to consider larger structural changes, including new tax revenue sources and new mechanisms for sharing AI-generated wealth. It points to possibilities such as basic income, sovereign wealth approaches, and equity-sharing models, while noting that the company is less certain about the right answers in that case.
Anthropic also emphasizes that companies deploying AI can choose to retrain workers instead of reducing headcount, but says voluntary action alone will not be enough without government standards.
Alongside the frameworks, Anthropic said it will commit $350 million to two initiatives. One is a $200 million Economic Futures Research Fund, which expands on an earlier program and is intended to support research trials and evaluations of public policy ideas. The other is a $150 million national fellowship program for early-career people working to bring AI benefits to communities across the country.
The company said the investments are meant to support policies and programs it is willing to help fund, including some that are not typically financed by private firms.
Anthropic framed the publication as an attempt to encourage planning before the pace of AI development overtakes government response. It said the economic transition is not predetermined and that choices made in the next few years will help determine whether AI broadens prosperity or concentrates it.
Although the economic proposal is focused on the United States, Anthropic said the principles could apply internationally and should be discussed with governments, including at the G7 and an upcoming AI summit in Geneva. The company also said it wants workers, unions, small business owners, and worker organizations to help shape the proposals.