OpenAI and Tech Giants Clash with California Over Controversial AI Safety Bill
OpenAI Opposes California’s AI Safety Bill, Citing Potential Industry Exodus
OpenAI, the artificial intelligence research laboratory, has voiced strong opposition to a proposed California bill that would impose strict safety protocols on AI models. The company argues that such legislation could drive AI firms out of the state, potentially hampering innovation in the sector.
The bill in question, SB 1047, introduced by California State Sen. Scott Wiener in February, aims to establish “common sense safety standards” for AI systems with development costs exceeding $100 million. Among its key provisions is the requirement for a “full shutdown” or kill switch for AI systems, designed to prevent “critical harms” such as cyberattacks or the development of weapons of mass destruction.
Jason Kwon, OpenAI’s chief strategy officer, warned that the bill could stifle progress and expressed concerns about regulating national security matters at the state level. “A patchwork of state laws is not the ideal approach for addressing these complex issues,” Kwon stated.
OpenAI is not alone in its opposition. Other Silicon Valley tech giants, including Meta and Anthropic, have also been lobbying against the bill. Meta cautioned that the legislation might discourage the open-source movement by exposing developers to significant legal liabilities. Rob Sherman, Meta’s vice president of policy and deputy chief privacy officer, emphasized that such regulation could hamper the broader tech ecosystem, as smaller businesses often rely on freely available models for innovation.
Anthropic, another AI company, has resisted the bill’s stringent preemptive regulations, advocating for a more balanced approach that wouldn’t impede progress in the field.
This is not the first time OpenAI has engaged in lobbying efforts against AI regulation. The company previously sought to ease regulatory requirements on general-purpose AI systems in the European Union’s AI Act, resulting in modifications to the final draft.
Despite industry pushback, Sen. Wiener defends the bill as “highly reasonable,” stating that it merely asks large AI labs to commit to practices they have already agreed to implement. The bill has already passed a vote in the state Senate and is scheduled for a final vote in the California Assembly at the end of the month.
As the debate intensifies, the outcome of this legislative effort could have far-reaching implications for the future of AI development and regulation in California and beyond. OpenAI and Sen. Wiener did not respond to requests for comment sent outside standard business hours.
