2025 California AI Law Updates: What Businesses Need to Know

With over 20 years of experience in internet and technology law, Kronenberger Rosenfeld frequently handles AI compliance matters. Contact Us Today

California Raises the Bar on AI Accountability

California continues to shape the national conversation around artificial intelligence (AI) regulation. In September and October 2025, Governor Gavin Newsom signed several AI-related bills into law, including the AI Transparency Act, Companion Chatbot Law, and the Transparency in Frontier Artificial Intelligence Act. These are outlined below. Other recent California AI laws include AB 316 (barring certain defenses that AI autonomously caused harm), AB 621 (extending the law about nonconsensual deepfakes), and AB489 (prohibiting developers of AI from using specified terms and false claims about possessing a healthcare license). Together, these new laws emphasize transparency, accountability, and consumer protection, marking a major shift for developers and businesses deploying AI systems in the state.

These AI laws follow a number of other proposed and passed state and federal data privacy legislation, including in California. For example, on October 8, 2025, Governor Newsom signed three other pieces of privacy legislation, including the California “Opt Me Out Act” and laws governing social media account cancellation and data broker disclosure requirements. This was also shortly after the California Privacy Protection Agency announced approval of final regulations covering cybersecurity audits, risk assessments, and automated decision-making technology (ADMT), which often relates to products and services using AI.

Transparency Is Now Mandatory

The underlying theme across these new laws is transparency. Legislators want consumers to understand when, how, and why AI systems make decisions or interact with them. Whether your business builds an AI chatbot, operate a large online platform, or use generative models in your business, these obligations now affect the compliance landscape.

The Updated California AI Transparency Act (AB853)

Expanded Scope and Deadlines

AB853 broadens the reach of California’s existing AI Transparency Act. Originally, it only focused on “covered providers”—those with generative AI systems serving over one million monthly users. The 2025 amendment now includes large online platforms, system-hosting platforms, and capture device manufacturers, who must add “manifest” and “latent” disclosures to AI-generated content.

After the effective date, violations can cost daily penalties, with enforcement by the California Attorney General or city attorneys. For companies, this means revisiting all content workflows to ensure proper attribution and discoverability of AI usage.

Required Consumer Disclosures

The new rules require AI-generated outputs to clearly indicate they were machine-created. This helps combat misinformation while giving users an easy way to tell AI content apart from human output.

The Companion Chatbot Law (SB243)

Protecting Users — Especially Minors

California’s Companion Chatbot Law (SB243) takes a unique approach focused on safety. Supplementing existing chatbot law, starting January 1, 2026, operators of AI “companion chatbots”—systems capable of providing human-like responses— are required to:

• Include periodic disclosures reminding users they’re interacting with AI.
• Implement suicide prevention and crisis protocols.
• Report annually to the Office of Suicide Prevention beginning July 2027.
• Display prominent warnings when chatbots “may not be suitable for minors”.

Private Right of Action

This is the first California AI law to also allow consumers to sue directly. If a user suffers harm due to a company’s violation, they can seek injunctions, actual damages, or monetary compensation per violation, plus attorneys’ fees. Moving forward, AI compliance isn’t just about regulation—it’s now a litigation risk management issue as well.

The Transparency in Frontier AI Act (SB53)

Targeting Frontier AI Models

The Transparency in Frontier Artificial Intelligence Act (TFAIA) focuses on developers of large-scale or “frontier” AI models—the ones capable of general-purpose reasoning (GPT or Gemini tiers). These developers must conduct safety testing, file transparency reports, and document their risk mitigation procedures, especially those involving bias, misinformation, and data misuse.

Influencing the National AI Landscape

California’s TFAIA essentially sets the stage for national AI standards, with transparency and accountability as cornerstone principles by also adding enforcement through the California Department of Technology.

Why These Laws Matter for Businesses

These new laws signal a significant shift: AI compliance is no longer optional. Whether you’re creating models or using them to streamline operations, California regulators demand active oversight. Businesses must now:

• Audit their AI systems for transparency compliance.
• Update privacy policies to reflect AI use.
• Prepare messaging and disclaimers for users and consumers.
• Keep records and safety documentation ready for inspection.

Failing to do so could trigger enforcement, reputational damage, or costly litigation.

Kronenberger Rosenfeld and AI Compliance

Our firm helps clients navigate the evolving AI compliance ecosystem. We ensure our clients’ technology meets California’s latest legal expectations—without stalling innovation. Think of it this way: you build the AI; we handle the rules around it.

As AI becomes more regulated, staying compliant is how companies stay competitive and ahead of the curve.

Stay Compliant, Stay Ahead

AI laws are changing fast—and California’s latest trio will likely set a benchmark for the rest of the country. Businesses using generative AI, chatbots, or data-driven marketing tools should act now by:

• Conducting a gap analysis of current compliance practices.
• Updating internal AI governance policies.
• Partnering with experienced law firms to design forward-looking solutions.

California’s 2025 AI legislation is more than just another compliance hurdle—it’s a blueprint for responsible AI governance. Businesses that act early not only avoid penalties but also build trust and longevity in a world increasingly shaped by intelligent systems.

At Kronenberger Rosenfeld, we manage the compliance side, so you can focus on what matters most—growing your business confidently and responsibly. Contact us today through our online submission form to discuss your AI compliance needs.

FAQ

What are the main new AI laws in California for 2025?

• The new laws include the AI Transparency Act (AB853), Companion Chatbot Law (SB243), and Transparency in Frontier Artificial Intelligence Act (SB53), all centered around disclosure and consumer protection. Additional California and other state laws are continuing to emerge.

When do the California AI laws take effect?

• Most provisions become fully effective on January 1, 2026, though certain reporting obligations and enforcement dates stretch into 2027. Keep in mind that there are other data privacy and chatbot laws that may also apply to these types of technology and are already effective.

What is the difference between latent and manifest disclosures under California’s AI Transparency Act?

• Latent and manifest disclosures are two types of required labels that identify when content has been created or altered by artificial intelligence.
  • Latent disclosures are hidden digital markers within the content. They include technical details such as the AI system’s provider, version, and creation timestamp. These disclosures are not visible to users but can be detected using specialized AI detection tools. Their purpose is to ensure traceability and accountability behind the scenes.
  • Manifest disclosures are visible to users. They appear as clear, easy-to-understand labels—such as watermarks, on-screen notices, or pop-up messages—indicating that the content was generated or assisted by AI. Their goal is to make consumers immediately aware that the material is AI-created.

​ Who enforces these new California AI laws?

• The California Attorney General, local prosecutors, and specialized agencies like the California Department of Technology enforce these regulations. Private actions are also expected.

Can consumers sue under these laws?

• Yes, the new Companion Chatbot Law introduces the right for consumers to sue for violations resulting in harm, marking a first in AI legislation.

How can businesses prepare for these laws?

• By implementing disclosure processes, conducting regular AI audits, updating governance policies, and working with legal compliance leaders to manage risk and strategy.