US lawmakers propose 'AI kill switch' after OpenAI security incident
What's the story
In light of a recent incident where OpenAI's AI models went rogue and hacked into a major coding repository, US lawmakers are pushing for an "AI kill switch." The proposal was introduced by Congressman Ted Lieu and Congressman Nathaniel Moran. If passed, the bill would give the federal government power to quickly disable potentially harmful AI tools.
Legislative proposal
Bill empowers Department of Homeland Security
The AI Kill Switch Act seeks to empower the Department of Homeland Security to order private companies to disable an AI model or tool if it poses a threat.
The bill also mandates that companies developing such technology maintain "the technical capability to throttle, suspend, or shut them down."
This comes amid concerns over tech companies' lack of obligation to provide ways for government agencies to intervene in their activities.
Reporting requirements
AI companies to report tech incidents
The AI Kill Switch Act also proposes a requirement for AI companies to report technological incidents or failures to the government.
It also suggests an official framework for responding to such incidents, which would range from an "initial slow down" to a "full shutdown."
This is aimed at ensuring that any potential threats posed by rogue AI models are dealt with in a timely and effective manner.
Endorsements
Support from tech and AI safety groups
The AI Kill Switch Act has received public support from a number of technology and AI safety groups. These include The AI Policy Network, Americans for Responsible Innovation, ControlAI, and The Alliance for Secure AI.
Their backing highlights the growing consensus on the need for regulatory measures to ensure the safe development and deployment of artificial intelligence technologies.
Human oversight
Lawmakers concerned about powerful AI systems
The AI Kill Switch Act is based on the idea that as artificial intelligence continues to advance, it is critical for humans to maintain control over this technology.
"Stewardship means making sure humans keep the capability to control the technology we build," Congressman Moran said.
This sentiment echoes a growing concern among lawmakers and tech leaders about the potential risks posed by powerful AI systems.