In 2026, a wide range of California laws regulating the development, marketing, and use of artificial intelligence (AI) go into effect. Together, these bills impose new requirements on generative AI developers, frontier-model companies, healthcare-related AI tools, platforms distributing AI-generated content, and businesses that rely on algorithmic pricing. With the deadline to comply coming up quickly, companies operating in California (or offering AI-enabled products or services to California residents) should assess how these laws apply to their technologies and update their governance and disclosure practices accordingly.
Most of the new obligations take effect in 2026, with some requirements extending into 2027 and 2028. Below is an overview of the key components of enacted bills AB 316, AB 325, AB 489, AB 621, and AB 2013.
AB 316: Liability for AI-Related Harms (Effective: January 1, 2026)
AB 316 applies broadly to any civil action where AI involvement is alleged to have caused damage. The bill limits affirmative defenses for civil liability, prohibiting defendants (which could include developers, modifiers, or users of AI) from raising an “autonomous-harm defense” in lawsuits alleging harm caused by AI-generated or AI-modified content. This defense, which might otherwise allow parties to shift blame to the technology’s independent decision-making, is explicitly barred to ensure human responsibility remains as AI models are deployed.
AB 325: Algorithmic Pricing and Antitrust (Effective: January 1, 2026)
AB 325 amends the Cartwright Act, a California anti-trust statute, to prohibit anticompetitive use or distribution of “common pricing algorithms.” A common pricing algorithm includes any methodology (computer-based or otherwise) that uses competitor data to recommend, align, stabilize, or influence prices or commercial terms. The statute creates two categories of liability: (i) use or distribution of a common pricing algorithm as part of a contract, combination, or conspiracy to restrain trade; and (ii) coercion to adopt an algorithm-recommended price or commercial term.
AB 489: Misleading Statements on Health Care Professional Oversight of Artificial Intelligence (Effective: January 1, 2026)
AB 489 prohibits developers and deployers of AI and generative AI technologies from using titles, terms, icons, post-nominal letters, or design elements that could falsely suggest the system is providing services from a licensed healthcare professional. Any implication — direct or subtle — that a licensed professional oversees the output is barred unless such oversight exists. This prohibition applies both to advertising and to in-product functionality for both AI and generative AI systems. Each misleading representation may constitute a separate offense, and state licensing boards are authorized to investigate and enforce violations, including through civil penalties.
AB 621: Expanded Protections Against Digitized Sexually Explicit Deepfakes (Effective: January 1, 2026)
AB 621 strengthens legal protections against non-consensual, sexually explicit “deepfakes.” The law broadens the definition of “digitized sexually explicit material,” clarifies that minors cannot consent to its creation or distribution and increases damages (up to $250,000 for malicious violations), and grants public prosecutors civil enforcement authority. These expanded remedies meaningfully increase the risks for individuals and entities involved in creating or distributing such material.
AB 2013: Mandatory Dataset Disclosure for Generative AI Developers (Compliance by: January 1, 2026)
AB 2013 requires developers of generative AI systems to publicly disclose detailed information about the datasets used to train their models. Disclosures must be posted on the developer’s website and updated when substantial system modifications occur. The law raises concerns from developers related to protection of intellectual property, confidentiality, and potential litigation exposure due to the breadth of required disclosures.
Takeaways to Prepare for Compliance
With multiple California AI-related statutes becoming enforceable in 2026 (and additional obligations later in 2027, and 2028 that will be covered in a subsequent post) companies need to assess the applicability of these laws now, which may include:
- Mapping and documenting training datasets now to ensure readiness for AB 2013.
- Evaluating the use of shared or third-party algorithmic pricing tools under AB 325.
- Reviewing their risk exposure related to the enabling of deepfake pornography under AB 621.
- Considering the allocation of liability in agreements given the prohibition on autonomous harm defenses by AB 316.
- Ensuring that their use of AI does not imply that their services are provided by a licensed healthcare professional unless overseen by a healthcare provider to comply with AB 489.
California’s AI regulatory framework is expanding rapidly. These new statutes are in addition to California’s regulations regarding Automated Decision-Making Technology, which were promulgated under the California Consumer Privacy Act and go into effect on January 1, 2026. Early preparation will help your company navigate the state’s emerging compliance landscape.