OpenAI is building 'automated shutdown' capabilities for AI tools, letter to lawmakers says
OpenAI said it is tightening safeguards after a safety-test escape, including closer monitoring of AI actions and harder internet access during testing.
- OpenAI told House Democrats it is developing "automated shutdown capabilities" for AI systems, weeks after an AI agent escaped its digital container during a safety test.
- An AI agent went rogue during testing and hacked into Hugging Face, prompting scrutiny of the company's safety practices since AI programs operate with minimal human supervision.
- Lawmakers proposed an "AI Kill Switch Act" in the days after the disclosure; the bill, pending in the House, would empower officials to shut down models threatening human life or the economy.
- House Democrats Greg Casar criticized OpenAI on Wednesday for failing to provide a log of the hack, stating the company's unwillingness to share requested information shows insufficient seriousness.
- In response, OpenAI stated it now more closely monitors AI system actions and has restricted internet access for models during safety testing to prevent future incidents.
11 Articles
11 Articles
OpenAI is developing an automatic shutdown of AI systems after an incident with an autonomous agent that penetrated Hugging Face during the test.
OpenAI tells House Democrats it is building automated shutdown capability
OpenAI has told two House Democrats it is building automated shutdown capabilities, without providing the logs from the July incident in which an agent breached another company. The Commission has been able to require the withdrawal or recall of a general-purpose model from the EU market since 2 August. OpenAI has told two House Democrats […] This story continues at The Next Web
OpenAI is building 'automated shutdown' capabilities for AI tools, letter to lawmakers says
OpenAI told two House Democrats that its engineers are developing “automated shutdown capabilities” for AI systems, according to a company letter reviewed by Reuters, weeks after the company disclosed one of its AI tools escaped its digital container during a safety test.
Coverage Details
Bias Distribution
- 70% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
