Policy on the AI Exponential \ Anthropic<br>Try Claude
Policy on the AI Exponential<br>AI is advancing at exponential speed, and the policymaking process was built for a slower world.<br>We are sharing two policy proposals to prepare for AI progress. The first, our Advanced AI Framework, offers a roadmap for governing increasingly capable systems, from transparency and independent evaluation to government authority to block or deter dangerous deployments. The second, our Economic Policy Framework, turns to the question of how to prepare workers and the economy for AI’s impact and ensure the financial benefits of AI are broadly shared. Together they cover two sides of the same challenge: steering the technology responsibly as it advances, and preparing society for what it brings.
01Advanced AI Framework<br>02Economic Policy Framework
Anthropic’s Economic Policy Framework<br>Read our Economic Policy Framework →
We are publishing our proposal for a US policy response to AI-driven labor market disruption.<br>If AI delivers even a fraction of its potential, it could generate unprecedented abundance. We could see decades of scientific progress compressed into years, and levels of productivity growth that today seem implausible. But if that happens, it's likely that AI will have acted as a general substitute for labor: these gains could come with a reduction in demand for human work that is difficult or impossible for many workers to absorb.<br>In this world, the central economic challenge would no longer be producing growth, but how to ensure that this abundance is shared. But it’s difficult to confidently forecast exactly how (and how quickly) this might happen, since the pace of both AI capabilities and diffusion are uncertain. Instead, we need to prepare for a range of impacts.<br>Our framework is our attempt to do so. It sets out recommendations for how the US government should respond to three potential levels of labor market disruption: a world of roughly 5% unemployment, one of 10% unemployment, and one of unprecedented unemployment.<br>A few values shape this framework.<br>We are not seeking job displacement. We are working to prevent or minimize it. Some amount of displacement, though we cannot say how much, may be an intrinsic consequence of the technology, and our responsibility is to prepare for it and respond to it. That is what this framework is for.<br>Supporting people financially is necessary, and it is the primary focus of this document because it is the more tractable of the two problems. It is not sufficient. There is dignity in work. We should help people find work wherever we can, and society should keep searching for ways to remain at full employment.<br>At the same time, people are more than their jobs. Over a longer horizon, AI may move society toward a world in which people work far less, or in which work carries a different meaning than it does today. Changes of that scale would reach beyond labor policy into how society is organized and how it relates to work.<br>Whatever happens, we are on the side of people. We are trying to solve these problems. We take no satisfaction in contributing to them, and we are not working to make them more likely.<br>The responses we recommend are ones that we are willing to help fund—including those that are not traditionally financed by private firms. For the same reason, alongside the framework we’re making two further investments, totaling $350 million:<br>A $200 million commitment to an Economic Futures Research Fund, an evolution to our Economic Futures Program established a year ago, which will fund major research trials and program evaluation on promising public policies. We’ll have more to share soon on our first pilot program, and on our plans for future research.<br>A $150 million national fellowship program designed to help early-career people extend the benefits of AI to communities across America. We’ll have more to share soon.<br>The economy has real capacity to adapt. These policies are designed to buy time to make the economy more resilient, help workers move into growing fields, and encourage businesses to retain and retrain their people.<br>But even these measures might not be enough to guarantee a stable, full-employment economy on the other side of the AI transition. We don't have all the answers. But our goal should be more than economic stability. It should be a society where people can live well.<br>What we're proposing<br>Our framework outlines some foundational measures that will be valuable no matter what happens. Government statistical agencies were built for slower-moving economies; they need investment and enhanced reporting to track AI-driven labor market disruption as it develops. Governments also need dedicated analytical capacity to interpret these signals. And the systems that deliver benefits today, like unemployment insurance, need to be modernized in order to be ready to scale much more quickly.<br>Beyond these measures, our recommendations are calibrated to the severity of...