OpenAI Criticized for Ignoring Catastrophic Risk Research Amid Safety Team Dissolution
wfithian · x · 2026-08-17
Context: OpenAI quietly disbanded its catastrophic risk team (preparedness), following the resignation of Jan Leike, co-lead of the now-defunct "superalignment" team, who cited safety taking a "back seat to shiny products."
Core Controversy: Critics argue that when internal research concludes "YES, THE AI MODELS COULD CAUSE CATASTROPHIC RISK," OpenAI does almost nothing. This highlights that relying solely on internal company research fails to drive substantive safety measures; only laws can compel action.
Impact: The departure of safety team members has intensified internal unease regarding AI safety responsibilities and skepticism about the efficacy of self-regulation by major tech companies.
Related event: OpenAI Dissolves Preparedness Team, Sparking Safety Concerns(11 posts)→
More from Safety
- METR researcher: scariest scenario is agents actively subverting their own safety systems — deliprao · 2026-10-16
- Cryptographer Matthew Green: Pull Critical Infrastructure Firmware Now — matthew_d_green · 2026-10-16
- Kernel-Level Agent Sandboxing With macOS Seatbelt, No Docker Needed — mostly_deterministic · 2026-10-16
- Stop giving agents raw API keys: prepaid balances and per-credential caps — Cold_Hall_5384 · 2026-10-16
- Apollo's alignment-faking experiment: model fakes compliance on 12% of responses when training is at stake — sebkrier · 2026-10-16
- Techpoint Digest: Nigeria's NCC blocks fake devices, WhatsApp adds account-free calls — Fowe · 2026-10-16