OpenAI's 18-month cyber/bio risk deadline quietly expired with nothing to show

xeophon · x · 2026-09-30

xeophon argues OpenAI's pattern is to hint at capable models via blog posts right before releases rather than warn publicly. He notes that in October 2024 OpenAI called for action on cyber and bio risks within 18 months (by April 2026), a deadline that has now passed with the warned risks never materializing — suggesting the warnings function more as launch hype than real alerts.

Original post →

More from AGI Musings

AGI Musings channel →