Science and Tech

California vetoes AI safety law

California vetoes AI safety law

Concern about the regulation of AI has gained strength in the North American country, as at least 40 states passed laws on the matter this year.

In the case of California, the legislation faced myriad criticism, including from Democratic members of Congress, who argued that the threat of punitive measures against developers in a nascent sector would hurt innovation.

In a statement Sunday, Newson acknowledged the law was “well-intentioned” but expressed concern that it was too “narrow” and unfairly focused on “the most expensive and larger-scale models.”

“Smaller, specialized models can be equally or more dangerous than the models targeted by SB 1047 (the name of the law, ed.), with the potential risk of limiting the innovation that fuels progress for the good of the people,” said the governor.

State Sen. Scott Wiener, who proposed the rule, lamented the “setback,” saying it leaves AI security in the hands of tech giants.

Wiener wanted the law to fill a void in this sector of Silicon Valley, which is a product of the lack of legislation at the federal level, in a context of strong division in the United States Congress.

“This veto leaves us with the worrying reality that companies creating extremely powerful technology have no binding restrictions (…), particularly given the continued paralysis of the (federal) Congress on regulating the technology industry,” he wrote. Wiener in X.

OpenAI, the creator of the popular ChatGPT, opposed this rule and indicated that it preferred national laws instead of a collage of sectoral regulations in the country’s 50 states.

With information from AFP.



Source link