Sacramento, California (AP) – Governor Gavin Newsom, Governor of California, signed a law on Monday to prevent people from using powerful artificial intelligence models to use possible catastrophic activities such as developing biological weapons or closing a bank system.
This step comes to Newsom to support California as a leader in regulation and criticized in omissions at a recent interview with former President Bill Clinton. The new law will establish some of the rules of the first state regarding large -scale AI models without compromising the state’s home industry, Newsom said. Thirty -two of the top 50 companies are in California and will have to comply with demands.
“California has proven that we can set rules to protect our communities, while ensuring that the growing AI industry continues to thrive. These laws get that balance,” the Newsom said in a statement.
Legislation requires PG companies to implement and disclose public security protocols to avoid the use of their most advanced models in order to cause serious damage. The rules are designed to cover AI systems if they meet the “border” threshold, which signals that they operate with a huge calculation power.
Such thresholds are based on how many computers do. Those who have prepared the rules have admitted that the sills of readings are an imperfect reference point to distinguish between today’s highest quality generating AI systems from the new generation, which may be even more powerful. Existing systems are mainly manufactured by California -based companies such as Anthropic, Google, Meta platform and Open.
Legislation describes catastrophic risks as what would cause at least $ 1 billion to damage or more than 50 injuries or deaths. It is designed to protect against AI use, which can cause mass disorders, such as hacking into the power grid.
Companies must also notify the state to any critical safety incidents within 15 days. The law creates the protection of informants to AI staff and creates a public cloud for investigators. This includes $ 1 million. USD fine for the offense.
For some technology companies, this contradicted, which stated that AI legislation should be executed at the federal level. However, the Anthropic said regulations help to smooth out the games and turn the official safety practice that many companies already voluntarily do.
“The developers can compete by ensuring that they remain transparent due to AI opportunities that lead to risks to public security, creating equal conditions where disclosure is mandatory and not selected,” the company said in its confirmation statement.
Signing will take place after the Newsom has been vetoing a broader version of legislation last year, which owns technology companies, who said the requirements were too inflexible and would be obstructed by innovation. Instead, Newsom asked a group of industry experts, including AI Pioneer Fei-Fei Li, to make recommendations on protective handrails around powerful AI models.
The new law includes recommendations and feedback from the Newsom AI Group and Industry Group, sponsors said. The legislation also does not submit the same level of reporting requirements for beginners to avoid the damage of innovations, said the state elder Scott Wiener from San Francisco, the author of the bill.
“As a result of this law, California is moving up again as a global leader in technology innovation and safety,” Wiener said in a statement.
Newsom’s decision is President Donald Trump in July announced a plan to abolish what his administration considers “heavy” rules to speed up AI innovation and strengthen the position of the US as a world leader. The Republicans of the Congress earlier this year have unsuccessfully attempted to ban states and areas to regulate a decade.
With no stronger federal rules, the states all over the country have spent the last few years trying to control this technology, fighting everything, starting with deepfacing elections to the AI therapy. In California, the legislator has accepted numerous bills of exchange this year to solve safety issues for AI talk programs for children and AI use at the workplace.
California was also an early acceptance of AI technology. The state, among other things, deployed the generous AI tools to spot fires and turn to the roadside congestion and road safety.
–
Associated Press reporter Matt O’Brien contributed to the message.