FPPC Debuts GenAI Use Policy, Plans to Expand Technology’s Use

The California Fair Political Practices Commission (FPPC) has unveiled its own agency-specific policy for generative AI (GenAI) use, building on top of existing guidance from the departments of Technology (CDT), General Services (DGS) and others.

The policy, in the draft stage, was debuted at a committee meeting the commission held Feb. 11. Adam Silver, FPPC chair, said at the meeting that the commission has plans in the works for implementing the technology.

“[This] marks the first formal step toward responsibly integrating the groundbreaking technology of AI in a way that strengthens our core mission — that’s promoting meaningful transparency, bolstering enforcement and increasing operational efficiency,” he said. “In the months ahead we will begin implementing tools that will allow us to better serve the public, our staff and the regulated community. I look forward to sharing more about those initiatives as they develop.”

Specifically, the FPPC has submitted a request to the Legislature for $710,000 to procure software to detect AI-generated content in campaign advertisements.

The new GenAI policy has five main pillars:

1. FPPC requires a “human in the loop” when using GenAI.
2. Information entered into GenAI systems may be subject to public records requests, and users should be able to document how they arrived at outputs. Information that can’t be released to the public shouldn’t be used.
3. GenAI tools not already approved by FPPC may require approvals before staff can use them. Microsoft Copilot is already approved and available for staff.
4. Staff must disclose to supervisors when they’ve used GenAI to complete work; public-facing work requires a specific disclosure.
5. The commission will provide training and working groups for staff related to GenAI.

Jesse Hidalgo, the commission’s CIO, told committee members that the policy builds on guidance from CDT, which issued a formal document for state agencies on the matter in 2024. DGS also maintains a list of policies related to AI across various sources such as the State Administrative Manual, the State Contracting Manual and the State Information Management Manual.

Hidalgo also mentioned CDT’s AI Community, where representatives from various state agencies bring forward use cases, brainstorm and share ideas.

FPPC is also working with peer jurisdictions across the country on AI, according to Silver.

“We already have a group of, I think at this point it’s about 15 or so different jurisdictions, who are working with us on basically brainstorming different ways that AI can be used in the ethics space,” he said. “We had our first meeting where we had representatives from Hawaii, New York, all over the country, who wanted to hear about what we’re planning to do here, the results of our first … meeting where we talked about detecting AI use in campaign ads.”

Silver said the commission will likely vote to formally approve the policy in March. But Hidalgo acknowledged that the fast-changing nature of the technology will necessitate regular review and updating of the policy.

“This is a very good first step, but we don’t expect it to be the last,” Hidalgo said.