본문으로 건너뛰기
All news

OpenAI Says It's Building an Automated Shutdown Capability, House Democrats Told

TL;DR: OpenAI has reportedly told House Democrats it is building an automated "shutdown" capability to halt its AI systems if problems arise.

Key points

  • OpenAI reportedly informed House Democrats that it is developing a safeguard to automatically stop its systems in the event of trouble.
  • The disclosure is said to have come in response to lawmakers' questions about AI safety and control.
  • The automated shutdown appears to function as a kind of "kill switch" for cases where an AI behaves dangerously or slips out of control.

Why it matters

It signals that AI-safety debates may be shifting from voluntary corporate pledges toward concrete technical measures. With U.S. lawmakers now pressing frontier AI companies for built-in control mechanisms, this offers an early read on where regulation could be headed.

Read more

뉴스레터 구독

무료 뉴스레터

매주 핵심 AI 소식, 한 번에 받기

쏟아지는 AI·LLM 뉴스 중 꼭 알아야 할 것만 골라 메일로 보내드려요. 뉴스레터 발송이 시작되면 구독자분들께 가장 먼저 보내드립니다.