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 like these. Many AI developers have since adopted policies built on this idea.

METR’s mission is to develop scientific methods to assess catastrophic risks stemming from AI systems’ autonomous capabilities and enable good decision-making about their development.

At some point, AI systems will probably be able to do most of what humans can do, including developing new technologies; starting businesses and making money; finding new cybersecurity exploits and fixes; and more. This could change the world quickly and drastically, with potential for both enormous good and enormous harm. Unfortunately, it’s hard to predict exactly when and how this might happen. Being able to measure the autonomous capabilities of AI systems will allow companies and policymakers to see when AI systems might have very wide-reaching impacts, and to focus their efforts on those high-stakes situations.

The stakes could become very high: it seems very plausible that advanced AI systems could pursue goals that are at odds with what humans want. This could be due to deliberate effort to cause chaos or happen despite the intention to only develop AI systems that are safe.¹ Further, given how quickly things could play out, we don’t think it’s good enough to wait and see whether things seem to be going very wrong. We need to be able to determine whether a given AI system carries significant risk of a global catastrophe.

We believe that the world needs an independent third-party that can scientifically and empirically study the capabilities and risks of AI systems. We also believe in the importance of transparency and openness: where possible, we publish all of our research and risk assessments, and seek to communicate our views to the wider world.

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. 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.