AI safety researcher slams reckless frontier model releases: we know how to turn it off, not align it

A safety researcher argued that since we don't know how to align AI—only how to monitor and shut it down—releasing frontier models without those safeguards is reckless, and listed indicators like model card detail, release delays, and pre-release testing with AISI.

2026-09-09 ~ 2026-09-09 · 3 related posts

• AI researcher argues OpenAI and Anthropic spend billions on safety while rivals race recklessly — Afinetheorem · 2026-09-09
• Model cards, release delays, and AISI pre-release cited as the real AI safety test — Afinetheorem · 2026-09-09
• We can't align AI but we can turn it off: researcher slams reckless frontier releases — Afinetheorem · 2026-09-09