California is at the forefront of AI governance with two laws that have the potential to shape the history of artificial intelligence: SB53 and SB243. The first would establish transparency rules for America’s fastest-growing industry, requiring companies to publicly cite safety policies and report important issues. Supporters hope the law will rein in an industry that often hides behind questionable disclosure practices, make it easier for companies and whistleblowers to report public safety issues, and instill concrete enforcement mechanisms to hold wrongdoers accountable.
The latter, on the other hand, seeks to protect vulnerable users from the negative effects of AI companion applications through age verification requirements, disclaimers, and self-harm prevention requirements. The legislation comes as debate over the governance of AI rages on across the country, with President Trump pledging to repeal federal laws that limit the field’s growth and Republicans in Congress seeking to bar state and local governments from scaling back the industry.
Meanwhile, New York state is considering its own Frontier AI Act, and several state legislatures are looking to regulate companion applications. Age verification laws, content bans, and even the possibility of VPN restrictions further highlight these ethical dilemmas, while the rapid growth of the industry requires urgent solutions. California is in a unique position in this debate. As the world’s fourth-largest economy, the Golden State is a leader in AI and is home to 32 of Forbes’ top 50 AI companies. According to Stanford University’s 2025 AI Index report, 15% of all AI job postings are in California, while more than 50% of venture capitalists’ AI investments went to Silicon Valley, according to Pitchbook research. This trend comes as no surprise since AI giants like Apple, Google, and Nvidia are based in the state.
Landmark transparency bill
The Frontier Artificial Intelligence Transparency Act instills fundamental guardrails for frontier artificial intelligence through “evidence-based policymaking” that balances transparency, security, and innovation considerations. The law establishes some basic parameters based on California’s March 2025 Working Report on the State of AI. To promote transparency, the bill requires developers to publish best practice policies on their websites to ensure companies comply with industry and international standards. The bill also creates a mechanism for holding companies accountable through civil penalties, while facilitating protections for whistleblowers.
To prevent large-scale disasters as AI becomes more pervasive in public and private infrastructure, companies must report potential safety hazards such as nuclear meltdowns, biological weapons, and cyberattacks to the California Department of Emergency Services. In addition to safeguards, Senate Bill 53 aims to foster public-private partnerships and foster industry development through a state-run computing cluster within the Government Operated Agency called CalCompute.
This bill is not the first time California lawmakers have attempted to govern AI. For example, Governor Newsom vetoed a stricter bill in 2024 that would require AI developers to include cutoff switches, cybersecurity protections, “substantial harm” safeguards, and pre-release safety testing. Critics say the new bill removes many of the regulatory elements of the 2024 bill and relies mostly on voluntary disclosures rather than security requirements that would hold companies accountable. But supporters tout the first-of-its-kind law as a blueprint for a federal framework. Meanwhile, industry reaction has been mixed. For example, security-minded Anthropic publicly supported the bill, while Meta and OpenAI initially campaigned against it but then acquiesced.
Make your AI companion safe
In October 2025, California enacted the nation’s most comprehensive safeguards for AI companions. Senate Bill 243 aims to protect young and vulnerable users by requiring suitability warnings, AI disclosure notices, and break reminders for minors. It also mandates new content protocols, preventing chatbots from creating suicide-related content and requiring companies to provide at-risk users with suicide prevention resources and referrals to crisis lines. Additionally, SB243 requires providers to prevent companions from sharing explicit content with minors, requires companies to publish these protocols on their websites, and gives users a “private right of action” to seek damages against chatbots. Starting in July 2027, AI companions will be required to submit annual reports to the California Department of Public Health detailing their response to user crises.
The law comes amid tensions over the influence of AI companions, following a series of lawsuits and investigations that have implicated companies such as Character.AI, Meta, and OpenAI in causing unwarranted harm or death to minors. As minors become increasingly dependent on technology, concerns about romantic relationships, inappropriate content, and misleading therapy bots are coming to the forefront. According to a study by Common Sense Media, most American teens use an AI companion, while the Center for Democracy and Technology found that more than 40% use the app for social advice and nearly 20% admit that they or someone they know has been in a romantic relationship with an AI companion.
This trend is likely to continue as AI companions proliferate across social media, with some AI companies even launching their own social media applications. As of 2025, five states have enacted mental health regulations regarding chatbots, but none have regulations as comprehensive as California’s. Whether other companies follow in Golden State’s footsteps could determine the direction of the nation’s fastest-growing industry.
