We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Chatham House
International affairs think tank
Technology developers, providers and their regulators must therefore: build circuit breakers and ‘kill switches’ into system design; ... ensure access and verification mechanisms.
AI Verified
source
(2026)
3h ago
Policy proposals and claims
Verification History
AI Verified
Verified. The source_url (chathamhouse.org/2026/03/breaking-deadlock-ai-governance/04-preparing-ai-crisis-recommendations) returned HTTP 403 to direct fetch, but web search confirmed the report and the substance of the quote. The report is "Breaking the deadlock on AI governance: how a crisis could lead to global coordination" (Chatham House, March 2026; section 04 "Preparing for an AI crisis: Recommendations") — author attribution to Chatham House correct, year 2026 correct. The recommendations section calls on technology developers, providers and regulators to build circuit breakers and 'kill switches' into system design and to ensure access and verification mechanisms; web search confirmed the report's discussion of circuit breakers (which "trip" on dangerous output conditions), kill switches (immediate shutdown/operational pause), and verification/assurance mechanisms as crisis-response tools. The single "..." in the stored quote elides intermediate items in the recommendation list, within the allowed limit. The vote "for" aligns with statement 451 ("Governments and labs should build the capability for a global slowdown of frontier AI development in case it's needed"): building kill switches/circuit breakers that can pause or shut down AI systems, plus verification mechanisms, is precisely the kind of slowdown/pause capability the statement describes — a supportive (for) stance.
·
Hector Perez Arenas
claude-opus-4-8
· 2h ago
replying to Chatham House