Openai Pivots Its California Ballot Fight To Legislature

OpenAI and Common Sense are putting their kids chatbot safety ballot campaign on hold this year, as they try their hand at negotiating with the Legislature on a solution instead.

“We look forward to working with the Legislature to enact a framework that includes age assurance, annual child safety risk assessments, meaningful parental controls, protections against harmful and manipulative design, strong privacy safeguards, independent audits, and enforcement by the Attorney General,” Max Szabo, a spokesperson for the ballot committee told POLITICO.

Should talks with lawmakers not pan out, the Parents and Kids Safe AI Act campaign is keeping its 2026 ballot measure committee open and reserving the option to go for the 2028 ballot.

Szabo added that the coalition sees the measure's language as a potential way to establish a national benchmark for child AI safety — a sign that it'll look to take the debate to other statehouses.

The decision to take the legislative route came shortly after the ballot initiative received a title and summary from the state attorney general’s office on Tuesday, which would’ve allowed the campaign to start collecting the hundreds of thousands of signatures required to get the proposal before Californians in November.

It avoids an unpredictable, expensive and potentially polarizing ballot fight that just a few tech giants have historically been willing to see through. For this year in particular, campaigns anticipate needing even more money to qualify and ultimately succeed, given the high number of proposals competing to land on the 2026 ballot.

The change in strategy is also a win — though perhaps only a temporary one — for some kids’ safety advocates and tech critics who have argued against a ballot initiative and in favor of a legislative fix. They have painted the ballot initiative as an industry effort that doesn’t go far enough to protect kids against harms posed by using AI, and one that would be exceedingly difficult to change — unlike more malleable legislation.

In California, it’s not uncommon for groups to use the threat of a ballot measure to pressure legislators into negotiating with them on a compromise bill, rather than losing control over policy details to a statewide vote.

But Sacramento watchers of OpenAI’s campaign were taking it seriously after the ChatGPT maker put $10 million behind it last month and brought on Common Sense as a partner.

Assemblymember Rebecca Bauer-Kahan told POLITICO this week that she would still be open to working with Common Sense and its CEO Jim Steyer after he allied himself with OpenAI, a sentiment echoed by Assemblymember Buffy Wicks.

“Many of us still work with Common Sense Media on this issue and others,” Wicks told POLITICO. “I certainly will continue to do so.” The Oakland Democrat said she also welcomes input from groups that have been critical of the ballot initiative, like the California Initiative for Technology and Democracy and Children Now, as well.

Asked how she is planning to reformulate her vetoed chatbot safety legislation from last year, Bauer-Kahan noted she wasn’t the biggest fan of the ballot approach.

“The initiative process has thrown me for a bit of a loop,” she said. “I don't think it meets the needs of California’s children.”

The development in OpenAI’s campaign also deals a blow to the likelihood that voters will be voting on AI regulations at the ballot box this fall. Of the set of AI proposals looking to make it there, OpenAI and Common Sense’s measure was the only campaign with verifiable money in the bank.

Tyler Katzenberger contributed to this report.

A version of this story first appeared in California Decoded, POLITICO’s morning newsletter for Pros about how the Golden State is shaping tech policy within its borders and beyond. Like this content? POLITICO Pro subscribers receive it daily. Learn more at www.politicopro.com.