Anthropic, once a defector from OpenAI, is now opposing legislation that would grant immunity to AI companies in cases of large-scale harm. The Illinois bill, SB 3444, has become a battleground between these leading US AI labs over how they should be regulated.

While Anthropic argues for more stringent liability measures, claiming that companies developing advanced AI models should bear some responsibility if their technology causes widespread societal harm, OpenAI defends the bill as a necessary step to ensure the safe deployment of frontier AI technologies. The debate reflects broader tensions in US AI policy and the role of regulation.

Despite the low likelihood of becoming law, SB 3444 has sparked significant lobbying efforts from both parties. Anthropic has been actively engaging with Illinois lawmakers, while OpenAI supports a harmonized approach at state level to inform federal legislation. Governor Pritzker’s office has also weighed in, expressing concerns about weakening legal accountability.

The crux of the disagreement lies in who should bear responsibility for AI-enabled disasters – a scenario that highlights the complex and evolving nature of AI regulation in the US. As these tech giants navigate their political strategies, the debate over liability and safety will continue to shape future policy discussions.