California gets new AI regulations, but not until 2025


California has shown it’s serious about regulating Big Tech in 2025 — and Big Tech has shown it’s serious about coming to the statehouse and fighting back.

The result was a series of laws designed to limit technological harms, but often in a watered-down form.

Take legislation from San Francisco Democratic Sen. Scott Weiner to prevent artificial intelligence systems from enabling catastrophes like biological weapons attacks. The the original version would be a must safeguards on artificial intelligence systems and possible liability imposed on their developers. Prav. Gavin Newsom vetoed it over concerns it would stifle innovation in the state’s burgeoning AI industry. This year’s version, signed into lawit simply requires major AI companies to publish safety frameworks and creates a pathway for reporting safety incidents.

Similar dilutions are seen in other technical offerings. on various data center regulation billsthose requiring disclosure of water and energy use failed, while one it simply allows regulators to look at those uses passed On AI chatbots that talk to childrena bill outright banning any harmful chats failed, while one simply required protocols for suicidal users became law. A underwent a similar process six bills to regulate algorithmic pricing to one signed by the governorprohibiting technology platforms from requiring their business customers to use their pricing recommendations.

This came after that another year of aggressive lobbying by tech companies, sometimes behind the scenes.

Still, advocates of more regulation won some decisive victories, including a new browser setting that prohibits websites from transferring personal data. Experts say that this “giving up” will eventually help US users

Meanwhile, California’s executive branch struggled with its defenses against online hackers, losing its top cybersecurity official amid discord in the office this position controls. Law enforcement agencies in the state also struggle to properly handle the digital data they collect, with many local police departments illegally sharing of vehicle traffic informationcollected by automated license plate readers, with federal agencies such as border patrol and immigration and customs control.

Outlook for 2026

Next year will see no end to the tension between protecting Californians from artificial intelligence and the impulse to protect the industry’s cash flow. An ambitious bill that would require disclosure of the use of AI in resulting decisions such as housing and education, will return. Data centers will be controversial as artificial intelligence will potentially increase energy consumption opening the door to nuclear power. Finally, the Trump administration and Republicans in Congress have drawn up proposals — so far not acted upon — to preempt state laws regulating AI. If passed, such plans would hit California the hardest.

Leave a Reply

Your email address will not be published. Required fields are marked *