Bonta sent the letter to the chief executives of Alphabet, Meta, Microsoft, OpenAI, Reddit, TikTok, X and YouTube on the day after the first televised debate between Vice President Kamala Harris and former President Donald Trump.
In his letter, Bonta urged each company to train experts to catch voter deception on their sites, and reminded the executives that the state's election code prohibits pushing out intentionally false information about voter eligibility or polling locations, as well as posting intentionally misleading media about a candidate within 60 days of an election.
Bonta also pointed out that state election laws prohibit voter intimidation and bribing people to vote or not vote a certain way.
"Your companies are core to the information ecosystem where many Californians get their news and guidance about elections, including the fundamental information about when, where, and how to vote," Bonta wrote. "Because your companies play such an important role in disseminating news and information, they are well-equipped to ensure that California voters that use your platforms, products, and services have access to accurate information about their right to vote."
In the letter, Bonta pointed to high-profile instances of AI interfering with elections, including in January when an AI-generated version of President Joe Biden's voice urged people not to vote in the New Hampshire Democratic primary.
"Troublingly," Bonta wrote, "many social media platforms have eased or eliminated their internal content-moderation policies, or reduced the resources dedicated to implementing those policies."
Bonta did not call out any particular site or service, although X fired thousands of employees including many on its content moderation teams, after Elon Musk bought the company in 2022. The site has seen a rise in right wing content since Musk's purchase of the company.
"We appreciate the Attorney General's concerns and look forward to continuing to communicate with his office as we address these challenges going forward," an X spokesperson said in an emailed statement the day after the letter came out.
Musk, who was not among the list of recipients, has openly endorsed Trump. Musk is the chief executive of AI startup, xAI, whose Grok bot is available through the X social media site.
The startup has drawn criticism for its Grok-2 image generator due to its readiness to depict political figures ranging from Trump to Democratic candidate Kamala Harris wielding weapons, doing drugs, or in other compromising positions.
Other AI image generators, such as those made by OpenAI, have stricter guardrails in place and are less willing to depict and discuss political figures and issues.
Alphabet, TikTok and OpenAI did not respond to emails requesting comment on the letter. Microsoft did not immediately provideda statement.
A Meta spokesperson referred to a post on how it labels AI-generated content on Facebook, Instagram and Threads. That policy from May says the company adds "AI info" labels "to a wider range of video, audio and image content when we detect industry standard AI image indicators or when people disclose that they're uploading AI-generated content."
OpenAI disclosed earlier this year how its chatbots were being used by state actors including Iran, Russia, Israel, and Chinese government-linked groups to spread falsehoods about worldwide conflicts and global politics. The company said at the time it had detected the efforts and taken steps to cut them off.
Reddit declined to comment, but referenced its content policy, which "prohibits AI-generated content presented to mislead, as well as many other forms of manipulated content, such as coordinated inauthentic behavior, falsified documents, and impersonation," and the posting of false information intended to stop people from voting.
Reddit also said in its election resources that it is using AI technology to flag harmful content. So far this year, the company said, it has not seen adverse effects on elections around the world from content posted on its site, but noted that it is adding more training for its safety teams.
Alphabet and Meta have been criticized for massive layoffs that also affected content moderation teams. And YouTube has reversed an earlier decision to remove content disputing the outcome of the 2020 election, allowing false claims that Trump won to remain online.
Meta has reportedly not fully enforced its policy around labeling political ads, and CEO Mark Zuckerberg recently wrote to Republican members of Congress saying the Biden White House pressured Facebook to remove some COVID-19 related content during the pandemic.
The California legislature has also acted on the issue of AI-driven political falsehoods. One bill, AB 2655, is awaiting a signature or veto from Gov. Gavin Newsom and would require big social media platforms to take down political deepfake photos and videos intended to mislead or confuse votes within a specified period leading up to elections.
Aside from Zuckerberg, the CEOs who received the letter from Bonta were Alphabet and Google CEO Sundar Pichai, Microsoft CEO Satya Nadella, OpenAI CEO Sam Altman, Reddit CEO Steve Huffman, TikTok CEO Shou Zi Chew, X CEO Linda Yaccarino, and YouTube CEO Neal Mohan.
© 2024 the San Francisco Chronicle. Distributed by Tribune Content Agency, LLC.
