CEOInterviews.AI
Start App
Try: AI strategy tariffs earnings guidance layoffs interest rates growth

Beth Barnes

Founder and Chief Executive Officer, Metr

Search every verified Beth Barnes interview, podcast appearance, and on-the-record quote — each transcript cross-checked by AI and human review to confirm speaker identity. Beth Barnes is the founder and CEO of METR (Model Evaluation & Threat Research), a research nonprofit she started in 2022 after leaving OpenAI. In recent appearances, she has discussed METR's work on evaluating AI capabilities and risks, including the development of tasks to measure how long AI models can work on real-world problems. She has stated that current models have roughly a 50% chance of completing tasks that take an expert human about one hour, with that capability doubling approximately every six months, though she has noted uncertainty in these estimates. Barnes has also emphasized the importance of independent, third-party evaluations and has called on frontier model developers to publish open-source evaluations to strengthen community oversight. Barnes has expressed concern about the lack of consensus on questions such as whether AI poses existential risks, and has argued that models can appear impressive on benchmarks while still failing at "messy" real-world tasks. She has said that AI could "totally transform the world, economically and socially," while also cautioning that current hype and actual future impact can coexist. Earlier in her career, she gave TEDx talks on effective altruism, advocating for prioritizing charitable work based on evidence of impact and discussing the expansion of moral consideration to future generations.

Recent appearances

  • The AI Progress Chart Everyone Is Misreading — Beth Barnes & David Rein

    May 04, 2026 · 📺 Machine Learning Street Talk

    Beth Barnes and David Rein on the one graph that ate the AI timelines discourse, and why the two people who built it are the most careful about how you read it. *SPONSOR* Prolific - Quality data. From real people. For faster breakthroughs. https://www.prolific.com/?utm_source=... Interview:    • AI training data will never be fully synth...   --- Beth Barnes and David Rein from METR on the one graph that ate the AI timelines discourse, and why the people who built it are the most careful about how it gets read. Beth founded METR after leaving OpenAI alignment. David is first author on GPQA…

    Watch on YouTube ↗

  • Measuring AI Capabilities, Trends and Risk (Beth Barnes & Lawrence Chan of METR) | The Curve 2025

    November 20, 2025 · 📺 Golden Gate Institute for AI

    Beth Barnes and Lawrence Chan (METR) go over what current AI systems can do in the real world, when to expect very capable systems, and why it will be hard to keep powerful models doing what we want. Along the way, they show how standard benchmarks often overstate real capabilities, highlight how performance collapses on “messy” real-world tasks, and demonstrate reward-hacking behaviors, arguing that careful, adversarial evals are essential on order to detect dangerous capabilities before they cause real harm. Now in its second year, The Curve is a conference that brings together a diverse se…

    Watch on YouTube ↗

  • The Most Important Graph in AI Right Now | Beth Barnes, CEO of METR

    June 02, 2025 · 📺 80,000 Hours

    AI models today have a 50% chance of successfully completing a task that would take an expert human one hour. 7 months ago, that number was roughly 30 minutes — and 7 months before that, 15 minutes. These are substantial, multi-step tasks requiring sustained focus: building web applications, conducting machine learning research, or solving complex programming challenges. Today’s guest, Beth Barnes, is CEO of METR (Model Evaluation & Threat Research) — the leading organisation measuring these capabilities. (Graph is discussed at 00:31:56 and visible here: https://metr.org/blog/2025-03-19-mea…

    Watch on YouTube ↗

  • Beth Barnes – METR Updates & Research Directions [Alignment Workshop]

    December 01, 2024 · 📺 FAR․AI

    Beth Barnes shares how METR is building scalable, interpretable metrics to reliably indicate when models become concerning. Barnes explains that although models are advancing in R&D tasks, challenges around cost profiles, elicitation, and transparency make it difficult to confidently set upper bounds on capabilities. She calls on frontier model developers to publish open-source evaluations to strengthen community oversight and mitigate emerging risks. Highlights: 🔹 Progress & Limitations – Models make gains on R&D tasks but are far from 1-day human performance. 🔹 Cost & Elicitation – Models h…

    Watch on YouTube ↗

  • 34 - AI Evaluations with Beth Barnes

    July 27, 2024 · 📺 AXRP

    How can we figure out if AIs are capable enough to pose a threat to humans? When should we make a big effort to mitigate risks of catastrophic AI misbehaviour? In this episode, I chat with Beth Barnes, founder of and head of research at METR, about these questions and more. Patreon:   / axrpodcast   Ko-fi: https://ko-fi.com/axrpodcast The transcript: https://axrp.net/episode/2024/07/28/e... Topics we discuss, and timestamps: 0:00:37 - What is METR? 0:02:44 - What is an "eval"? 0:14:42 - How good are evals? 0:37:25 - Are models showing their full capabilities? 0:53:25 - Evaluating alignment 1…

    Watch on YouTube ↗

  • METR's Talk on Evaluations Research - Beth Barnes, Daniel Ziegler, Ted Suzman

    April 12, 2024 · 📺 METR (Model Evaluation & Threat Research)

    Hey everyone uh I'm Beth and I'm going to be talking about U meter's research on evaluations so this will be three parts uh first I'm ...

    Watch on YouTube ↗

  • Evaluating Language Models for Autonomous Capabilities - Beth Barnes

    March 25, 2024 · 📺 Apart - Safe AI

    Beth Barnes, CEO of METR, gives a keynote presentation on the creation of evaluation tasks for AI agents as part of the Code Red Apart Hackathon, hosted by Apart Research [https://apartresearch.com/] and METR [https://metr.org] in March of 2024. Join us online for the live Q&A and project presentations at   / discord   Learn more about the hackathon ⭢ https://apartresearch.com/event/codered Previously at Deepmind, OpenAI, and the Alignment Research Center (ARC), Beth recently founded Model Evaluation and Threat Research (METR) as a neutral organization to evaluate advanced AI systems for sa…

    Watch on YouTube ↗

  • 6 - Debate and Imitative Generalization with Beth Barnes

    April 08, 2021 · 📺 AXRP

    One proposal to train AIs that can be useful is to have ML models debate each other about the answer to a human-provided question, where the human judges which side has won. In this episode, I talk with Beth Barnes about her thoughts on the pros and cons of this strategy, what she learned from seeing how humans behaved in debate protocols, and how a technique called imitative generalization can augment debate. Those who are already quite familiar with the basic proposal might want to skip past the explanation of debate to 13:00, "what problems does it solve and does it not solve". Link to B…

    Watch on YouTube ↗

  • Expanding circles | Beth Barnes | TEDxYouth@Tallinn

    April 19, 2018 · 📺 TEDx Talks

    Our values change with time. Therefore it is even more important to take the long perspective and think what would be normal for future generations. How can we act so that the next generations wouldn't be embarrassed for what we did? Beth is a computer science student at the University of Cambridge. She has been involved with Effective Altruism, a growing social movement based on the idea of using evidence to figure out how to do the most good, founding an EA society at her high school. She believes that steering the development of technologies such as synthetic biology and artificial intell…

    Watch on YouTube ↗

  • Effective Altruism | Beth Barnes | TEDxExeter

    June 09, 2015 · 📺 TEDx Talks

    ------ Effective altruism is a growing social movement founded on the desire to make the world as good a place as it can be, the use of evidence and reason to find out how to do so, and the audacity to actually try. This video explains some of the core ideas of effective altruism and discusses some of the ways we can fix the world's biggest problems. Read more here: http://www.effectivealtruism.org/ http://www.effectivealtruism.org/ TEDxExeter 2015 took the long view both back into the past and ahead into the future. We asked our speakers to help us understand the challenges that face us now…

    Watch on YouTube ↗

The summary and quote tags on this profile are produced with AI assistance from verified, first-person interview transcripts, then checked by our team to confirm the speaker's identity and the accuracy of every quote. See how we verify →