A new OpenAI-backed coalition is working to advance AI kids’ chatbot safety legislation in Sacramento this year.
With more than a dozen kids safety advocates, community groups and other organizations on board and recruitment efforts continuing, the Kids and Parents Safe AI Coalition aims to pass what it hopes is the nation’s strongest child safety AI law.
The group, which is first being reported in Decoded, is urging legislators in Sacramento to enact some of the key tenets of the chatbot and AI safety initiative that OpenAI’s ballot committee pivoted away from last month.
“This is one of the strongest efforts of its kind in the United States,” said Tecoy Porter, president of the civil rights group National Action Network’s Sacramento chapter, in a statement.
Organizations already signed onto the coalition include California LULAC, Youth Beat, the National Diversity Coalition, Healthy Black Families Inc., Social Equity California and others.
“California must act now to establish clear standards that protect children and ensure companies are designing these technologies with kids’ safety in mind,” said Erin Hogeboom, director of San Diego for Every Child, which has also signed onto the coalition, in a statement.
Emails went out last week to groups about the coalition, with more than one coming from former Republican Assemblymember Jordan Cunningham, who along with Democratic Assemblymember Buffy Wicks co-authored the Age-Appropriate Design Code. Cunningham is part of the team working on building the coalition, the group confirmed.
The new effort comes after OpenAI and other tech companies closely negotiated last year’s SB 53 AI safety measure that Gov. Gavin Newsom signed into law, but each company largely brought their own concerns to the table and did not operate with coalitions.
Earlier this year, OpenAI teamed up with Jim Steyer’s online kids’ safety group Common Sense Media on the ballot initiative, putting $10 million into a ballot committee before deciding to focus on a legislative push. That committee remains active.
Not everyone in the kids safety advocacy world is signing on to the new group.
Despite their previous collaboration, Common Sense was not listed among the groups that are part of the coalition. The nonprofit declined to comment.
Common Sense and OpenAI’s ballot alliance elicited backlash from some kids’ safety and civil society groups, with some circulating letters of opposition to the effort. What followed was a sit-down with Common Sense’s head of AI Bruce Reed last month where some groups expressed feelings of mistrust and even betrayal. Some pointed out that a ballot initiative would be difficult to amend, and would lock in standards they felt could fall short of protecting kids online.
The initiative was dropped not long afterwards
There are two major chatbot bills in the Legislature, one sponsored by Assemblymembers Rebecca Bauer-Kahan and Wicks, and the other moving in tandem from state Sen. Steve Padilla.
The authors previously told Decoded those yet-to-be fleshed out measures will likely focus on age verification, third-party auditing, parental controls and preventing the sale of kids’ data, among other safeguards, and will not be sponsored by any outside groups, like Common Sense, which is often onboard with major kids’ safety legislation.
The new OpenAI-backed coalition has put out a set of principles on AI safety that closely track those of the ballot initiative but hasn’t said which bills it might engage on.
What is clear from the coalition’s website is the effort will focus on similar issues as those outlined by Wicks, Padilla and Bauer-Kahan, although amendments have yet to go in print for those bills.
The coalition said it is focusing on urging the Legislature to implement online age assurance and parental controls, as well as protecting kids from inappropriate content and protecting their data and shielding them from targeted ads, along with risk assessments and safety audits. It will also advocate for safeguards to keep systems from creating emotional dependence, and holding AI companies accountable for protecting children through California attorney general’s office investigations and financial penalties.
Similar to the ballot measure team-up with Common Sense, not every kids’ safety group is happy about OpenAI-backed coalition-building efforts.
“We have no plans to join a coalition with OpenAI,” said Josh Golin, executive director of kids safety nonprofit Fairplay, in a statement.
Fairplay was among the groups that received emails seeking to engage them on the effort.
“If OpenAI wants to help, it can do so by not interfering with proposed legislation that will actually hold it accountable for the harm it is causing,” Golin added.