OSTP Director Kratsios: if you think your AI is unsafe, just stop building it
_NathanCalvin · x · 2026-09-21
OSTP Director Michael Kratsios commented on the AI pacing debate, echoing VP Vance's earlier line ("if you are building Frankenstein, stop"). His position: if you believe the technology you're developing is unsafe and don't want it out in the world, you can just stop it — no one needs to force you. He pushed back on company leaders who claim they'll adopt safety measures and say the government must act while continuing to build anyway, saying such companies should either answer this question better or consider whether they can actually slow down more. Context: Fox News interview following Trump's announcement of an AI Force. Safety researcher Nathan Calvin flagged the comments as notable.
More from Safety
- NIST's CAISI publishes assessment of Z.ai's GLM-5.3 cyber capabilities — 233C · 2026-09-21
- 16 Critical Breaks Hit NGCC Post-Quantum Candidates; Researcher Launches Own Forum — jedisct1 · 2026-09-21
- AI product privacy: derived data and behavioral data need protection too — goyalshaliniuk · 2026-09-21
- 5 Types of Data Every AI App Should Protect, From PII to Behavioral Signals — goyalshaliniuk · 2026-09-21
- Welfare and alignment are the same problem: curiosity without stakes has no corrective loop — habitante · 2026-09-21
- US Treasury Secretary says US proposed an AI national-security notification mechanism to China — ShakeelHashim · 2026-09-21