Ex-OpenAI staff tell UK Parliament they faced pressure to downplay AI risks
connoraxiotes · x · 2026-10-17
- Former OpenAI employees Steven Gadler and Daniel Kokotajlo testified before a UK House of Commons committee that they faced implicit and explicit pressure to minimize frank talk of AI risks at OpenAI — countering narratives that labs hype AI risk.
- Kokotajlo recalled executives saying that once people understood what OpenAI was building, they wouldn't let the company do it.
- Someone also suggested copying "TikTok's playbook" — using OpenAI's platform to get "tons of our users and influencers to go promote the benefits of our technology."
More from AGI Musings
- threepointone's hunch: mech interp is all code, so alignment could be solved like debugging — threepointone · 2026-10-17
- Grady Booch: what matters isn't matrix math or ion migration, it's what the system does — Grady_Booch · 2026-10-17
- The web's death is inevitable: models are commodities and agents must negotiate API access — _AustinCalvert_ · 2026-10-17
- Terry Tao's blog runs satirical guest post slamming AI-for-math hype as an IPO marketing gimmick — DanielWhiteson · 2026-10-17
- Builder argues alignment-to-operator beats alignment-to-creator, names two biggest risks — willcb · 2026-10-17
- OpenAI's mass math-solutions dump slammed for not warning mathematicians first — RileyRalmuto · 2026-10-17