--- id: "2026-10-03-geoffrey-irving-time-50-percent-pause" url: "https://postcutoff.com/e/2026-10-03-geoffrey-irving-time-50-percent-pause/" as_of: "2026-11-19T22:11:00+01:00" date: "2026-10-03" date_precision: day category: policy-safety importance: 3 confidence: high status: [Confirmed] sources: 3 editor: human editor human_review: null version: "2026-10-07" --- As of: 2026-11-19 22:11 CET. Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/ Canonical page: https://postcutoff.com/e/2026-10-03-geoffrey-irving-time-50-percent-pause/ # Ex-UK AISI chief scientist Geoffrey Irving puts the chance AI kills everyone at ~50% Full title: Ex-UK AISI chief scientist Geoffrey Irving puts the chance AI kills everyone at ~50% and calls for stopping frontier AI development now (TIME) In a TIME op-ed on Oct 3, 2026, Geoffrey Irving, former chief scientist of the UK AI Security Institute and now co-founder and chief scientist of Resolution, wrote: "I believe there's about a 50% chance we all die because of the development of smarter-than-human AI systems." He argued that key scientific questions will not be settled until it is too late and that "we can, and should, stop frontier AI development immediately". PA Media paired the piece with David Robinson's OpenAI resignation, giving it mainstream pickup. ## Key facts - Irving's estimate: 'about a 50% chance we all die' from smarter-than-human AI; he says the number is not precise - Capabilities he names as the route to catastrophe: hacking, persuasion, hiding its thoughts from humans, and planning/coordination between agents - Unresolved questions he says will only resolve too late: whether skills transfer, whether AI exceeds human level, whether small-scale misbehaviour scales with capability, whether current safety methods work on superintelligent systems - Call to action: stop frontier AI scaling now, with international cooperation modelled on nuclear non-proliferation; he notes US support for a pause 'has increased tremendously over the past few weeks' - His X thread announcing the op-ed had ~14k views by Oct 5 (fxtwitter); the PA story ran in the Irish Examiner and BreakingNews.ie on Oct 4 - Background: Irving previously worked at OpenAI, Google DeepMind and Google Brain (his X bio) ## What happened Irving, one of the best-known technical AI-safety researchers to have led a government safety institute, argued in TIME that waiting for scientific certainty about AI risk means waiting until it is too late, and called for an immediate halt to frontier development. ## Why it matters It adds a former government chief scientist to the people publicly calling for a pause in autumn 2026, during the same week as David Robinson's resignation essay and amid rising US poll support for a pause. It is an opinion piece, not new evidence. ## Your AI and this story - GPT-6.2 Astra (training cutoff May 2026): 125 days after its cutoff - Claude Opus 5.5 (training cutoff June 2026): 95 days after its cutoff - Gemini 4 Hydrogen (training cutoff March 2026): 186 days after its cutoff - Grok 4.7 (training cutoff May 2026): 125 days after its cutoff ## Sources 1. [TIME (Geoffrey Irving): We Won't Know the Answers to AI's Most Important Questions Until It's Too Late](https://time.com/article/2026/10/03/we-won-t-know-the-answers-to-ai-s-most-important-questions-until-its-too-late/) (time.com, official) 2. [Geoffrey Irving on X: 'we should stop frontier AI scaling now… I wrote about this for TIME'](https://x.com/geoffreyirving/status/2106392345327567274) (x.com, official) 3. [Irish Examiner (PA): OpenAI safety leader quits amid further expert warning of '50% chance we all die'](https://www.irishexaminer.com/news/arid-41919463.html) (irishexaminer.com, press) ## Changes - 2026-10-05 (filed): Created after the PA pickup ([We Won't Know the Answers to AI's Most Important Questions Until It's Too Late](https://postcutoff.com/p/2026-10-03-geoffrey-irving-time-too-late/)) ## Related - 2026-10-03: [Departing OpenAI safety lead David Robinson writes in The Atlantic](https://postcutoff.com/e/2026-10-03-david-robinson-atlantic-openai-culture-broken/index.md) - 2026-09-23: [Sanders and Casar introduce the Ban Artificial Superintelligence Act, with a pause on advanced AI and a new Department of AI](https://postcutoff.com/e/2026-09-23-ban-artificial-superintelligence-act/index.md) - People: [David Robinson](https://postcutoff.com/person/david-robinson/), [Geoffrey Irving](https://postcutoff.com/person/geoffrey-irving/) - Archived: 2 posts and 2 videos, listed in https://postcutoff.com/e/2026-10-03-geoffrey-irving-time-50-percent-pause/index.json