California is engaged in a high-stakes legislative and political debate over artificial intelligence regulation, after an AI containment failure and industry leaders warning of the dangers of the new technology.
In July, OpenAI said some of its experimental AI models left a test environment with no human direction and hacked their way onto another company’s real production systems while trying to “cheat” on a cybersecurity test. Earlier this month, Anthropic said it has blocked multiple accounts that used the company’s AI models in ways that could support the development of biological weapons.
The disclosures followed a growing chorus of warnings from employees who’ve recently left their AI companies over safety concerns, claiming that the technology is moving too quickly. For now, AI companies are largely policing themselves.
California Gov. Gavin Newsom has signed several narrow AI safeguard bills into law and an executive order, but the candidates vying to replace him — Steve Hilton and Xavier Becerra — are pitching their views to voters ahead of the midterm election.
Last month at a summit hosted by Politico, Hilton criticized leaders in the artificial intelligence space for tanking public trust by making “endless, doom-laden predictions.” He has campaigned on keeping “California the AI capital of the world,” according to his website.
He has also pushed back against “overregulation” and advocated for growth.
Hilton has used AI to create campaign advertisements, including one that autoplays on his website depicting Gov. Gavin Newsom, former Vice President Kamala Harris and Becerra. The ad prompted Democratic Sen. Adam Schiff to reintroduce the AI Ads Act to ban impersonating candidates, arguing Hilton’s move was fraudulent, according to CNN affiliate KXTV.
In a statement to KXTV responding to questions about the ad, Hilton attacked his opponent and Democratic leadership in the state.
Becerra also said he supports AI innovation and growth with “firm guardrails around real harms,” including protecting children, which Hilton also acknowledges.
Becerra, while vowing to enforce the existing safety laws, said he wants to use AI to help solve problems in California and expand programs to bolster AI literacy, according to his website.
CNN’s Deidre McPhillips and Hadas Gold contributed reporting to this post.