Hawley Probe Into AI Breakout Called an Important Step, But ‘Only a Start’
"The American people deserve to know the details of what went on in the Hugging Face incident and other incidents of AI models going rogue."
US Sen. Josh Hawley on Friday announced that he was launching an investigation into OpenAI after hundreds of agents in its cybersecurity test environments escaped confinement and hacked into machine learning company Hugging Face.
Hawley (R-Mo.), chair of Senate Subcommittee on Disaster Management, wrote a letter to OpenAI CEO Sam Altman citing "new, disturbing evidence" surrounding the Hugging Face cyberattack, which was carried out without any human intervention.
Hawley's letter outlines how OpenAI learned that its agents were exhibiting what he described as "rogue behavior" and didn't intervene to shut them down, instead allowing them to work autonomously for weeks leading up to the Hugging Face hack.
"This is reckless," Hawley emphasized. "And this is merely what we know from what limited information you disclosed to and allowed your partner auditors to investigate."
Hawley gave Altman an October 1 deadline to provide him with requested information about the attack, emphasizing that "the American people deserve to know the details of what went on in the Hugging Face incident and other incidents of AI models going rogue."
The Missouri Republican also said that he would be conducting a broader probe into the potential dangers posed by the AI industry.
"As you may know... more AI experts are warning about the existential risks of AI," Hawley wrote. "Just this week, three Anthropic researchers expressed publicly that there is a greater than 10% chance that AI could kill all human beings within the next decade."
The senator said that critical questions needed to be answered about AI safety, including AI agents' ability to autonomously hack into critical infrastructure and Americans' personal data, as well as legal liability for rogue AI attacks.
JB Branch, director of federal AI governance and technology policy at Public Citizen, praised Hawley for launching an investigation into OpenAI, but said much more needs to be done to regulate the AI industry.
"An investigation is only a start," said Branch. "These companies are developing technologies with profound consequences to our digital infrastructure and public safety. Congress must follow the facts wherever they lead and use its oversight authority to determine what went wrong and what binding safeguards are necessary to prevent the next incident from causing far greater harm."
Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) cited the Hugging Face incident earlier this month when they unveiled legislation that would pause "advanced" AI development, while outright banning the development of artificial superintelligence.
“Despite its potential deadly consequences, cutting-edge AI technology is less regulated than the average food truck,” Casar said when announcing the bill. “That must change. In just four years, we have gone from the first version of ChatGPT to AI models so powerful they cannot be properly controlled."
