Bri Treece

New @Fathom_org report: “Who Watches the Watchers” How do we build a governance infrastructure for AI that actually works, that keeps pace with the technology, produces meaningful safety outcomes, and scales beyond what governments can do alone? We brought that question to a full-day workshop at the International Association for Safe and Ethical AI (@IASEAIorg) conference in Paris. 🧵

Every AI lab watching this now understands the implicit message: comply fully, or be made an example of. This the most consequential relationships of our generation. Today's decision makes it dramatically harder to get that relationship right.

The costs to Anthropic will be significant. But the deeper damage is not to any single firm. It is to the broader ecosystem of trust that must exist if frontier AI and democratic governance are going to coexist responsibly.

We at Fathom are deeply troubled by reports that the federal government is designating Anthropic a "supply chain risk" for refusing to renegotiate contract terms that prohibit the use of its technology in autonomous weapons systems or the surveillance of Americans.

Final day. @BriTreece21 takes the stage to kick off Day 3 of The Ashby Workshops 2026 presented by @Fathom_org. The question: Which societal-level challenges raised by AI will require our greatest collective action to solve? One final push to turn conversations into action.

Worth reading. From Davos, @AndrewFATHOM maps the gap between AI certainty on the surface and uncertainty in the rooms where decisions get made, including for workers and communities. Governance is how we close it.

We just published our vision for the AI Century. At @Fathom_org, we’ve been asking the same questions many of you have: ➡️ What kind of future are we building? ➡️ What type of world will our children inherit? ➡️ Are we doing enough? ➡️ Can we even do anything at all? This essay is our answer, grounded in historical precedent, economic data, and a clear-eyed view of what’s coming. AI is unlike any technological revolution that came before, and it's moving faster than any transition in history. But here's what history also teaches us: we've been here before. Every industrial revolution brought dislocation, unrest, and uncertainty – until we rebuilt the social contract and reestablished an equilibrium. Labor protections. Social insurance. Public education. None of it was inevitable. People built it. That's what Fathom is for. We're developing the policies, institutions, and frameworks to make this transition work for everyone. The window for action is open. We intend to seize it.

@jackclarkSF Congrats on the new arrival! And I'm impressed you can read complete sentences at this moment!

Can't think of a better way to read There Is No Antimemetics Division than in the hallucinatory polyphasic sleep state that a newborn entails. I doze near the crib, change the baby, read a few pages, both sleep, wake up unsure if I'm dreaming, read more, etc. Perfect!

AI’s failures around teen self-harm demand better oversight. Our latest blog explores how Independent Verification Organizations (IVOs) could help prevent tragedies and create a safer, better future with AI. Read the full explainer here 👇 #AI #AIGovernance #Safety

Our co-founder and Chief Strategy Officer @AndrewFATHOM on America's AI Action Plan: "The Action Plan’s call to build an AI evaluation ecosystem is critical for increasing public trust and adoption of AI." Full statement below.

1/ 🔥 AI agents are reaching a breakthrough moment in cybersecurity. In our latest work: 🔓 CyberGym: AI agents discovered 15 zero-days in major open-source projects 💰 BountyBench: AI agents solved real-world bug bounty tasks worth tens of thousands of dollars 🤖 Autonomously. A pivotal shift is underway — AI agents can now autonomously do what only elite human hackers could before.

I view SB 813 as a step forward—particularly in fostering innovation in AI safety through the creation of independent multi-stakeholder regulatory organizations (MROs). It would also set in motion crucial work to establish the legal infrastructure and adaptive standards needed to keep pace with the rapid evolution of AI. However, to build on this momentum, I believe certain changes are essential, including: 1⃣Set up robust safeguards to ensure that MROs remain truly independent from market pressures, along with the possibility of public comments on the methodology proposed and published by each MRO. 2⃣Develop clear liability insurance requirements for developers who opt out of the certification process, along with reporting and transparency requirements to enable effective liability lawsuits if needed. 3⃣Affirm that AI developers can be held liable if their systems cause catastrophic harm that could have been plausibly anticipated and thus could have been prevented. 4⃣Restrict the protection from liability so that it is not a total and blanket protection but only covers ordinary negligence, not gross negligence, to keep an incentive for AI developers to improve their safety methodology. 5⃣Find ways to ensure state oversight of standards proposed by each MRO, including the creation of an independent expert panel to review their plans, public comments and set outcome-based safety principles and targets.

@deanwball @mkratsios47 Congrats! Super exciting!

I am happy to announce that I have joined the White House Office of Science and Technology Policy as a Senior Policy Advisor on AI and Emerging Technology. It is a thrill and honor to serve my country in this role and work alongside the tremendous team @mkratsios47 has built.

Don't miss @deanwball's latest on why AI liability is the next policy frontier.

RT @Fathom_org: The Ashby Report is here! Dive into this year’s Ashby Workshops, exploring AI, governance, and the future of policy. And d…

Governing AI globally is a maze right now. European leaders are breaking ranks, there are power vacuums everywhere, and people are looking to the Trump Administration for guidance. However, the question is not who will lead, but what new model of governance can rise to meet this