California Senate Bill 53

California Senate Bill 53 is a state law that requires companies building very large, powerful AI systems to publicly disclose their safety plans and report incidents when something goes wrong. Think of it like a large chemical plant that has to file its safety procedures with regulators in advance and report immediately if an accident happens. Instead of just trusting companies to police themselves, the law forces a baseline level of transparency.

The law is one of California's attempts to set standards on its own, at a time when there is still no unified federal AI regulation in the United States. Large AI companies like OpenAI initially pushed back against this kind of regulation because of the burden it created, but depending on public opinion and the political climate, some have shifted their stance and even asked for the law to be made stronger.

In news coverage, this shows up in headlines like "OpenAI, which once opposed California's AI safety law, now asks for it to be strengthened" — describing how companies initially resisted California's frontier model legislation before later changing their position. It's important not to confuse this law, which regulates the safety and transparency of large AI models themselves, with SB 1119, a separate bill focused on protecting minors.