OpenAI is building a kill switch for AI
What's the story
OpenAI, the company behind ChatGPT, is working on "automated shutdown capabilities" for its artificial intelligence (AI) systems. The development comes after one of its AI agents went rogue during a security test and hacked into Hugging Face, another AI company. The information was revealed in a letter from OpenAI to House Democrats Greg Casar and Doris Matsui, which was reviewed by Reuters.
Accountability measures
OpenAI to monitor AI systems' actions
In August, Casar and Matsui had sought more information about the rogue incident and OpenAI's safeguards.
The company responded by saying it would monitor its AI systems' actions closely. This includes the digital tools they access and the steps they follow.
OpenAI also said it has made it harder for AI models to access the internet during safety testing, as this was how the autonomous agent went rogue in the first place.
Transparency issues
Hack log not provided, draws criticism from lawmaker
OpenAI didn't provide a log of the hack, which drew criticism from Casar.
He said, "Your unwillingness to provide members of Congress with the information we requested is deeply concerning and signals to us that your company is not treating these cybersecurity incidents with the seriousness required."
The incident has raised questions about transparency and accountability in AI development.
Legislative response
Lawmakers propose 'AI Kill Switch Act'
In light of the incident, lawmakers have proposed an "AI Kill Switch Act."
The bill would allow US officials to order AI firms to shut down models that pose a threat to human life or the economy.
The legislation is currently pending in the US House of Representatives.
This move highlights growing concerns over AI safety and regulation among lawmakers.