OpenAI is supporting a proposed California bill, SB 1047, which would establish a legal framework for advanced AI developers. The bill would create a ‘state of the art’ legal defense, shielding companies from liability for harms caused by their AI models if they followed established safety protocols at the time of development. This protection would extend to potential catastrophic risks, including mass casualties or financial disasters. Proponents argue the measure is necessary to foster innovation by providing legal certainty, while critics warn it could effectively grant immunity for severe harms and reduce incentives for rigorous safety testing. The legislation also proposes establishing a new state agency to oversee powerful AI systems and a mechanism for whistleblowers to report unsafe models. The full article details the debate surrounding this significant proposed shift in AI liability law. Read the full article at: https://www.wired.com/story/openai-backs-bill-exempt-ai-firms-model-harm-lawsuits/