California Advances Four AI Chatbot Regulation Bills Through Committee

--- headline: "California Advances Four AI Chatbot Regulation Bills Through Committee" slug: california-ai-chatbot-bills-committee category: policy story_number: "14" date: 2026-04-28 ---

# California Advances Four AI Chatbot Regulation Bills Through Committee

Sacramento pushed ten artificial intelligence bills through legislative committees last week in the most concentrated burst of AI lawmaking any single state has produced this year, positioning California to once again set the terms of the national debate over how to govern machines that talk back.

Four of those measures target chatbots directly. Three address AI in healthcare. Two regulate AI in employment decisions. And one tackles the thorny question of training-data provenance. Together, they form the broadest simultaneous committee-stage advance of AI legislation in the 2026 session, and they arrive just as a Cooley analysis warns that the patchwork of state-level AI regulation is reaching an inflection point.

The Chatbot Bills

The four chatbot bills that cleared committees each attack a different facet of human-AI interaction. AB 1988, the Preventing AI User Self Endangerment (PAUSE) Act, passed out of the Assembly Health Committee. It targets the growing body of evidence that AI chatbots can encourage self-harm, particularly among vulnerable users. AB 2023, focused on companion chatbot safety for children, cleared the Assembly Privacy and Consumer Protection Committee. Its Senate companion, SB 1119, passed the Senate Judiciary Committee. AB 1609, which regulates customer service chatbots, advanced through the Assembly Judiciary Committee.

The children-focused bills carry particular urgency. "When a conversation between children and a chatbot goes wrong, the consequences can be dire," Senator Steve Padilla said when introducing the companion measures. "We have seen the consequences of our inaction towards the dangers posed by social media, and the stakes are too high to make the same mistakes again."

Assemblymember Rebecca Bauer-Kahan, a co-author of AB 2023, was more pointed. "What happened to Adam Raine is not an edge case; it is a warning," she said. "These platforms are designed to create dependency, and not a single one has been held accountable for its harmful failures. AB 2023 and SB 1119 change that, with real protections, real enforcement, and a clear message that California will not wait."

Healthcare, Employment, and Provenance

The healthcare cluster is equally aggressive. SB 903, which regulates the use of AI and the transcription of patient information by mental health professionals, passed the Senate Privacy, Digital Technologies and Consumer Protection Committee. AB 2575, addressing AI in healthcare services more broadly, was amended and cleared the Assembly Privacy and Consumer Protection Committee. AB 1979, also amended, passed the same committee.

On the employment front, SB 947, covering automated decision systems in workplaces, was amended and passed the Senate Privacy, Digital Technologies, and Consumer Protection Committee. AB 2027, which establishes prohibitions on certain uses of worker data by AI systems, cleared the Assembly Privacy and Consumer Protection Committee. Both bills target the use of AI in hiring, firing, promotion, compensation, and displacement decisions, areas where algorithmic bias has been documented repeatedly.

The provenance bill, AB 2713, known as the California AI Transparency Act, addresses system provenance data and requires disclosures about data used to train AI systems. It was voted out of the Assembly Privacy and Consumer Protection Committee and is now on a third floor reading, making it the furthest along of the ten measures.

A National Regulatory Landscape at an Inflection Point

California is not legislating in a vacuum. According to MultiState.ai, state lawmakers in 45 states had introduced 1,561 AI-related bills as of March 2026. ComplianceHub declared 2026 "the year of the chatbot bill" after tracking over 300 AI-related measures introduced in just the first month of the session. An AI2Work analysis counted 78 chatbot safety bills across 27 states.

Cooley, in an April 24 analysis titled "State AI Laws: Where Are They Now?," described the US regulatory landscape as being "at an inflection point," noting that hundreds of proposed state measures have created a fast-developing patchwork while federal action threatens to reshape or constrain state-level initiatives. The White House released its National Policy Framework for Artificial Intelligence in March, urging Congress to enact sweeping AI legislation to preempt certain state AI laws that risk "stifling innovation."

California Governor Gavin Newsom has moved on a parallel track, issuing Executive Order N-5-26 on March 30, directing state agencies to draft recommendations for AI safety requirements for companies doing business with the state. The executive order covers illegal content, bias, and civil rights protections.

Building on SB 243

The new bills build on California Senate Bill 243, the companion chatbot law that took effect January 1, 2026, making California the first state to impose safety, governance, and reporting requirements on chatbot operators. SB 243 requires disclosures when users could be misled into thinking they are interacting with a human, mandates crisis referral notifications, and imposes special protections for minors including periodic reminders that they are chatting with AI.

The ten bills advancing through committees now represent a second wave, extending the regulatory framework from disclosure requirements into substantive restrictions on how AI systems can operate in healthcare, employment, and customer-facing contexts. Colorado, Connecticut, Florida, and Hawaii are all moving their own AI bills simultaneously, but no state matches California in sheer legislative volume or the breadth of sectors covered.

For AI companies, the compliance calculus is getting more complex by the week. With committee votes behind them, these bills now face floor votes and potential amendments, but the direction of travel in Sacramento is unmistakable: California intends to regulate AI not as a single technology but as a collection of specific risks, one bill at a time.