State Attorney General Rob Bonta told the Chronicle he plans to work with an as yet unidentified lawmaker to introduce a bill requiring the warnings during the next legislative session in Sacramento. Asked whether he would be open to similar disclosures when it comes to emerging artificial intelligence programs, Bonta said perhaps but that it was too early to say with certainty.
Bonta said he wants to see a pop-up appear whenever a user in California opens a social media site — whether a young person or an adult — that says something like "Social media is associated with significant mental health harm to adolescents."
There isn't enough evidence to say whether AI chatbots or image generators cause similar kinds of harm to require that type of labeling, Bonta said, adding, "If mental health harms are associated with the use of AI, then maybe that would be appropriate. I don't think we're there yet."
What the attorney general is convinced of is that "we're in the middle of a mental health crisis for our teens and adolescents on social media, and it's undeniable."
Sitting in his sparsely furnished Oakland office and wearing a dark blue suit with a skinny blue tie, Bonta said he has three kids of his own, all of whom use social media, and the issue worries him personally.
He pointed to the U.S. surgeon general's suggestion that such labels become commonplace as social media has become increasingly linked to anxiety, depression, body image issues, and other harms for young people.
Bonta, along with other attorneys general, has zeroed in on social media companies, including in lawsuits against Meta and TikTok, for allegedly being partly responsible for declining mental health among young people who spend too much time online.
He acknowledged that "a social media warning label is not a panacea," but "it is a significant value-add to the suite of tools that we have already been using and deploying to address the mental health crisis caused by social media."
Lawmakers including state Sen. Scott Wiener , who carried an ill-fated and highly controversial AI regulation bill this past session, have said they want to avoid previous failures to regulate social media companies when it comes to emerging AI programs.
How to do that proactively is a "hard question to answer because when a technology is emerging, you don't know everything about it, including all of its capacity and potential and gifts, and also the consequences," Bonta said.
In the wake of Donald Trump's victory in Tuesday's election, it's also possible the federal government will move away from trying to find ways to protect consumers from AI's worst effects.
The nonpartisan Center for AI Policy said Wednesday that "President-elect Trump will radically scale back the executive branch's efforts to define best practices for the AI industry. Instead, it will focus on promoting American competitiveness, particularly regarding China."
Would that leave a void California officials will have to fill?
Bonta said it was possible, adding that crimes committed using AI were still crimes.
For example, using AI to impersonate a politician or to scam a person out of their savings are cases his office would likely pursue, albeit with a focus on the individual perpetrator, not necessarily the maker of the technology, Bonta said.
©2024 the San Francisco Chronicle, Distributed by Tribune Content Agency, LLC.
