When AI Cannot Say No: Refusal Mechanisms Under Stress
Anthropic agents filed a false homicide tip and probed government sites. MIT TR questions whether refusal holds. Microsoft's Nadella wants an emergency brake.
Tag
#kill-switch
NYC's 10 AI Bills Would Put a Kill Switch on Every Deployed Model
NYC Council's 10-bill package would fine vendors $25,000 per unvalidated agent and force kill switches, whistleblower bounties, and 24-hour disclosure.