We run AI safety training and research programmes including LASR Labs, AI Governance Taskforce, and the Arcadia Alignment Team.
London
Joined May 2026
- Tweets17
- Following18
- Followers348
- Likes7
Arcadia Impact retweeted
New blogpost! We surveyed 52 AI safety professionals (spanning 20+ orgs) on how AI tools accelerate them.
We find that the median reported speedup is 1.58x. This is substantially lower than similar figures reported by Anthropic and OpenAI researchers.
🧵👇
Arcadia Impact retweeted
New blog post! 🚨
Our team recently built a research scaffold to try to speed up our work. In this post we discuss what we found, and how we’re planning on studying automated alignment research going forward. lesswrong.com/posts/zGaQ3SS9…
Arcadia Impact retweeted
🚨 New paper on alignment midtraining!
We stress test alignment midtraining (AMT) at scale. Via systematic ablations, we highlight novel failure modes and identify many critical implementation details required for good performance. We open-source our work.
🧵 👇
AI Governance Taskforce Autumn 2026 applications are open – deadline Monday 31st August 2026.
For senior or mid-career professionals transitioning into AI governance, focused on mitigating catastrophic risk.
What this involves:
👥Applied research in small teams
📄Policy-relevant outputs (papers, briefs)
🌎Fully remote, global, part-time (8 hrs/week), 12 weeks total
💼Selective, voluntary research fellowship
🕑Deadline: Monday 31st Aug 2026
📅Cohort dates: 12 Oct 2026 - 15 Jan 2027
More info and apply: arcadiaimpact.org/ai-governa…
Arcadia Impact retweeted
Research update from the scalable oversight team at Arcadia Alignment!
Debate is a proposed training method for aligning models on tasks we don't have labels for. We want to study this empirically: does it work on today's models? lesswrong.com/posts/hb8pv3zy…
Arcadia Impact retweeted
New blogpost from the alignment team @ArcadiaImpact !
TL;DR your model organisms might be fried, like Spongebob
🧵 👇
*NEW* AI alignment research team!
We're announcing the new alignment team @ArcadiaImpact. A London-based team, working closely with @AISecurityInst to tackle 3 ambitious agendas in AI alignment!
👇 🧵
Putting together this dream team wasn't easy. Over the course of 3 months spinning up, we learned some valuable lessons about starting new AI safety organisations - read about that here!
lesswrong.com/posts/4onALBND…
We're also hiring! If you'd enjoy ambitious & impactful work, this could be your chance!
Programme Manager: arcadiaimpact.org/alignment-…
Ops & Comms Associate: arcadiaimpact.org/ops-and-co…
Deadline 23rd June