State AI Laws in Effect: A Practical Guide to the U.S. AI Regulatory Landscape
Overview with reference-level detail per law
Technical-legal, concise and tabular
0 0 5
A regulatory tracker from Orrick's AI Law Center listing U.S. state AI laws in effect or with future effective dates, organized by state and AI scope (e.g., AI CSAM, deepfakes, political advertising, healthcare, automated decision-making, user-facing AI, transparency). Each entry gives the relevant law, link, effective date, key requirements, and enforcement penalties. The page is a legal reference rather than a tool tutorial.
main points
unique insights
practical applications
key topics
key insights
learning outcomes
• main points
1
Comprehensive, structured coverage of 278 state-level AI law entries across many jurisdictions and AI scopes
2
Clear tabular organization with effective dates, key requirements, and enforcement penalties for each law
3
Filterable by state and AI scope, with links to primary legislative sources and recent amendments noted
• unique insights
1
Highlights a trend toward 'AI Transparency' and provenance requirements (e.g., California AI Transparency Act's latent disclosures and capture device obligations) that may affect developers and platforms
2
Shows emerging regulatory categories such as AI auditor registries and Independent Verification Organizations, signaling a growing compliance services market
• practical applications
Useful for legal, compliance, and product teams needing to understand which U.S. state AI laws apply to their products and when obligations take effect. It is a reference for regulatory awareness rather than a hands-on AI tool learning resource.
• key topics
1
State-level AI laws and effective dates in the U.S.
2
AI-generated deepfakes, CSAM, and intimate image regulations
3
AI transparency, provenance, and training data disclosure requirements
• key insights
1
Consolidated, continuously updated view of fragmented state-level AI legislation
2
Side-by-side comparison of scope, requirements, and penalties across states
3
Organized by AI regulatory category for quick cross-jurisdictional research
• learning outcomes
1
Identify which U.S. states have enacted AI-specific laws and their effective dates
2
Understand categories of AI regulation such as deepfakes, transparency, healthcare AI, and automated decision-making
3
Recognize compliance obligations and enforcement penalties relevant to AI product deployment
As artificial intelligence becomes embedded in everyday business operations, a growing patchwork of state laws is reshaping how companies develop, deploy, and disclose AI systems. Unlike a single federal framework, these state-level rules vary in scope, timing, and penalties, making it difficult for businesses to know which obligations apply to them. Understanding which states have AI laws in effect today is therefore a practical compliance necessity, not merely an academic exercise.
“ How the U.S. AI Law Tracker Works
The U.S. AI Law Tracker provides a consolidated view of key AI laws across states that have enacted them. Each entry identifies the AI scope the law addresses, the relevant statute, a link to the source, its effective date, the principal requirements it imposes, and the enforcement mechanisms and penalties available. Users can filter by state or by topic, such as AI deepfakes, AI in employment, or AI privacy. The information is refreshed monthly and reflects only states with defined AI laws.
“ AI-Generated Content, Deepfakes, and Election Integrity
Several states have moved quickly to address synthetic media. Alabama prohibits distributing materially deceptive AI-generated media intended to influence elections, while offering a safe harbor where a clear disclaimer is provided. Arizona has enacted both a general deepfake law, granting individuals a right to seek declaratory and injunctive relief against unconsented digital impersonations, and a political deepfake law restricting deceptive synthetic media of candidates within 90 days of an election unless a clear disclosure is included. California requires committees producing qualified political advertisements to disclose AI generation or substantial alteration. Penalties range from misdemeanor and felony charges to civil damages and per-violation fines.
“ AI in Healthcare: Utilization Review and Patient Communications
Healthcare is an area of particularly active regulation. California's AI Healthcare Utilization Law requires health care service plans and disability insurers using AI tools for utilization review to ensure determinations are based on specified information and are fairly and equitably applied. A separate California provision requires health facilities, clinics, and physician offices using generative AI for patient communications about clinical information to include a disclaimer that the content was AI-generated and to give patients clear instructions for reaching a human provider. Communications reviewed by a licensed or certified provider are exempt from this disclosure requirement.
“ Transparency, Training Data, and Content Provenance Requirements
Transparency obligations are expanding. California's Artificial Intelligence Training Data Transparency Act requires generative AI developers to publish summaries of training datasets, including sources, data point counts, whether copyrighted or licensed material is included, and collection time periods, beginning in 2026. The California AI Transparency Act, as amended, requires providers of covered generative AI systems to embed detectable latent disclosures in AI-generated image, audio, and video content and to offer detection tools. Starting in 2027, large online platforms must detect and surface provenance data, and from 2028, capture device manufacturers must embed latent disclosures by default. Violations carry penalties of up to $5,000 each.
“ Companion Chatbots, Minors, and User-Facing AI Rules
User-facing AI faces some of the most targeted rules. California's Companion Chatbot law requires operators to notify users when a chatbot is artificial and not human where a reasonable person might be misled, to conduct and document risk assessments focused on minors, to commission independent child safety audits, and to maintain protocols addressing suicidal ideation and self-harm content. A related measure, Adam's Law, adds independent child safety audits and penalties of up to $15,000 for intentional violations, with a private right of action for children harmed by certain violations. California has also proposed a temporary prohibition on toys containing AI companion chatbots. Earlier, the California Bot Act prohibited deceptive bot communications intended to incentivize commercial transactions or influence elections, subject to a disclosure safe harbor.
“ Automated Decision-Making, Privacy, and Algorithmic Accountability
Automated decision-making is increasingly addressed through privacy frameworks. Alabama's Personal Data Protection Act, effective in 2027, gives consumers the right to opt out of solely automated processing used to make significant decisions about credit, housing, insurance, education, employment, health care, and access to basic necessities. Penalties can reach $15,000 per violation. Businesses operating across states should map these opt-out rights against sector-specific rules and other state privacy statutes to build consistent consumer-facing processes.
“ AI Auditors and Independent Verification
A new category of regulation establishes oversight of the auditors themselves. California's Artificial Intelligence Auditors law requires the Government Operations Agency to maintain an AI Auditor Registry, prohibits unregistered individuals from conducting covered audits required by state law, and imposes formatting, record-keeping, and advertising standards on registered auditors. Separately, California's Artificial Intelligence Safety Independent Verification Organizations framework allows qualified assessors to be designated by the State to evaluate risks posed by AI systems and models, with designation subject to revocation. These measures signal a movement toward formalized third-party verification of AI systems.
“ Ownership of AI Models and Generated Content
Ownership questions are addressed directly in some jurisdictions. Arkansas sets default rules under which the person providing input or directives to a generative AI tool owns the generated content, provided it does not infringe existing rights. The person supplying training data owns the resulting model if the data was lawfully acquired and no contrary agreement exists. Where an employee generates content or trains models as part of their job, the employer is deemed the owner. These defaults can be overridden by contract, making agreements with employees, vendors, and data providers especially important.
“ Compliance Strategies and Staying Current
Because state AI laws differ in effective dates, definitions, and penalties, businesses should inventory their AI use cases, identify the states where they operate or where their users reside, and map each system against applicable requirements. Key steps include reviewing disclosure obligations, establishing opt-out and human-review pathways, documenting risk assessments, maintaining training data records, and monitoring legislative developments. Since the landscape changes frequently, subscribing to regular updates and consulting with counsel for jurisdiction-specific guidance will help organizations remain compliant while continuing to innovate.
We use cookies that are essential for our site to work. To improve our site, we would like to use additional cookies to help us understand how visitors use it, measure traffic to our site from social media platforms and to personalise your experience. Some of the cookies that we use are provided by third parties. To accept all cookies click ‘Accept’. To reject all optional cookies click ‘Reject’.
Comment(0)