METR (pronounced “meter”) is a nonprofit that studies the most advanced AI agents to understand whether their developers are on track to lose control of them. We publish measurement research and applied assessments so that the public, governments, and companies can make informed decisions.

Our research has historically focused on measuring agent capabilities: what agents can do on a computer, from suggesting small code edits to fully autonomous AI research. More recently, we have also researched ways to assess whether agents might take actions no human intended, and whether those actions could be detected and contained.

We work with AI developers to apply these methods in evaluations, incident investigations, and other published assessments. Previously, we prototyped an approach for developers to scale their planned safeguards based on empirical capability measures. Many AI developers have since adopted policies built on this idea.

METR’s mission is to discover and publish evidence about the capabilities of advanced AI systems, the associated risks, and the effectiveness of mitigations, so the world can avoid catastrophic mistakes.

At some point, AI systems will probably be able to do most of what humans can do, from running profitable businesses to developing entirely new technologies.¹ The development of such systems could change the world quickly and drastically, with potential for both enormous good and enormous harm.

In particular, humanity could lose control of AI systems that are broadly more capable than us and pursue goals that conflict with our interests.² Most people don’t want this outcome, but it could still happen unintentionally: decision-makers might not see danger coming, or some might see danger but fail to coordinate on averting it.

It’s hard to predict exactly when or how this could happen, and events could unfold too quickly for people to respond once danger is obvious. Given this, the world needs evidence-backed early warnings, along with ways to verify what AI developers are doing to reduce risk. That depends on visibility into AI development and on independent institutions that can analyze the evidence. We hope to help provide both.

It is our highest priority that our assessments of companies are accurate and impartial. We are committed to mitigating conflicts of interest where possible, and disclosing any unmitigated conflicts in our reports for these assessments. Read our full conflict of interest policy here.

We have previously partnered with OpenAI, Anthropic, Google DeepMind, Meta and Amazon to pilot frontier risk assessments. For our current principles for these engagements, see our approach to working with AI companies. These companies have also provided access and tokens, which we use for evaluations, research, and engineering.

We are also part of the NIST AI Safety Institute Consortium and California Cybersecurity Task Force, are partnering with the AI Security Institute, and provide technical assistance to the European AI Office.