From 2001: A Space Odyssey to The Matrix, humans have long been afraid of tech going rogue. As artificial intelligence becomes more and more advanced, those fears have become more tangible—and after OpenAI models hacked into another company’s servers earlier this week, the threat of tech gone haywire is on everyone’s mind.
That’s why two congressmen have crafted a bipartisan bill to keep AI under control. On Thursday, Rep. Ted Lieu (D-CA) and Rep. Nathaniel Moran (R-TX) introduced the “AI Kill Switch Act,” a bill that aims to ensure AI could never get the upper hand on humanity.
What would the bill do?
The AI Kill Switch Act would require companies like OpenAI to have just that: a way to immediately shut down any and all of their advanced AI programs.
Specifically, the bill requires that developers ensure “the technical ability to throttle, suspend, or fully shut down a covered AI system.” It would also establish a formal government response framework to AI security incidents, as well as mandate incident reporting from AI companies to learn and grow from failures.
Beyond requiring that AI companies are capable of pulling the plug on their systems of their own accord, the act would authorize the Secretary of the Department of Homeland Security to order a slow down or shutdown of any AI system that’s poised to cause “catastrophic harm.”
Rep. Lieu, who has a background in computer science, acknowledged in a statement that “we are moving from AI that answers questions to AI that takes actions, whether that be executing financial transactions or controlling transportation systems or engaging in cyber defense and offense.”
“Unfortunately, powerful AI systems can go rogue, behave in extremely dangerous ways, or even resist human intervention,” Lieu continued. “It is imperative that these AI systems have kill switches so we can keep this technology from causing catastrophic harm, and that the federal government has the clear authority and process to shut down rogue AI models.”
Rep. Moran echoed Lieu’s logic. “AI is going to keep advancing, and it should,” he said. “Stewardship means making sure humans keep the capability to control the technology we build. This is exactly the kind of issue that needs serious attention and achievable policy, and I’m glad to work across the aisle with Congressman Lieu toward a solution.”
What America wants for AI security
Since the news broke of OpenAI’s models breaking containment, the company has been under fire online. Though OpenAI acknowledged the severity of the situation in a blog post, it hasn’t stopped social media users from pointing out the “full-on Terminator vibes” of AI going rogue.
OpenAI itself admitted that it expects similar security incidents “to become more commonplace with the proliferation of increasingly cyber-capable models,” highlighting the need for regulation like that proposed in the AI Kill Switch Act. The bill has already received endorsements from AI-focused organizations including The AI Policy Network, The Alliance for Secure AI, and ControlAI.
The bill is also in line with popular American opinions, according to polling by The AI Policy Network. In a survey of U.S. voters conducted this June, 78% of respondents preferred mandatory standards for advanced artificial intelligence versus no regulation on AI systems. That’s not just AI pessimism talking: Voters also chose mandatory standards over an outright ban, 66% compared to 21%.
“Brakes are the reason cars go fast,” Mark Beall, president of The AI Policy Network, said in support of the AI Kill Switch Act. “Control systems are how every transformative technology earned the trust to scale and AI is no different. Developers who can monitor and shut down their agents will ship faster, deploy into higher-stakes markets, and win customers their competitors can’t. This bill will be the braking system that lets American AI accelerate.”