Arizona AI Laws and Regulation (2026)

Arizona regulates artificial intelligence through targeted, sector-specific laws rather than a comprehensive framework. Enacted statutes include HB 2394 (A.R.S. § 16-1023), which creates a civil action for election and other digital impersonations; SB 1359 (A.R.S. § 16-1024), which requires an AI disclosure on candidate deepfakes distributed within 90 days of an election; SB 1295, which criminalizes AI-generated voice, image, or video impersonation done with intent to defraud as a Class 5 felony; and HB 2175, which bars health insurers from using AI as the sole basis for claim denials.

Overview of Arizona AI Laws

Arizona has emerged as one of the more active states in AI regulation. Unlike many states still deliberating their first AI bill, Arizona has already enacted several targeted AI laws addressing deepfakes in elections, AI-generated voice fraud, child exploitation imagery, and healthcare claim denials.

The state's approach combines legislative action with administrative policy. Governor Katie Hobbs signed multiple AI bills into law during the 2024 and 2025 sessions, while the Arizona Department of Administration published a comprehensive Generative AI Policy governing state agency use of the technology.

Arizona's 2026 legislative session continues to address AI regulation, with new bills targeting chatbot safety for minors and AI education. Meanwhile, the Attorney General's office has taken enforcement action against AI platforms generating harmful content.

This article covers Arizona's enacted and pending AI legislation, executive actions, and enforcement activities. The enacted statutes described here were re-checked against the Arizona Revised Statutes in September 2026. Bill status for the 2026 legislative session is stated as of March 2026 and may have moved since. You should consult an attorney for advice specific to your situation.

Enacted Arizona AI Laws

Arizona has passed five significant AI-related laws since 2024. Each targets a specific category of AI harm rather than attempting comprehensive regulation.

HB 2394: Election Deepfake Protection (A.R.S. § 16-1023)

Governor Hobbs signed House Bill 2394 into law on May 21, 2024, creating Arizona's first law specifically addressing AI-generated deepfakes. Notably, the bill itself was partially drafted using AI technology, which may be a first in state legislation.

Who the law protects:

The statute lets a candidate for public office or political party office, or any Arizona citizen, bring a civil action over deepfake media that impersonates them. To win preliminary declaratory relief, subsection F requires the plaintiff to prove the basic elements plus at least one additional showing, including that the person is:

• Depicted nude or in a sexual act
• Depicted committing a crime
• Reasonably expected to suffer personal or financial hardship
• Likely to suffer irreparable reputational damage

Political candidate protections apply specifically during the 180-day period before a scheduled election for that office.

Enforcement mechanism:

Rather than imposing criminal penalties, HB 2394 creates a civil enforcement framework, and the default remedy is narrower than it is often described. Subsection A states that "the sole remedy on this cause of action is preliminary and permanent declaratory relief except as otherwise expressly provided by this section." A declaratory judgment is a court ruling that the media is a digital impersonation. It is not money.

Injunctive relief and damages sit in subsection I, and they are available only if the plaintiff proves the subsection A elements plus all three of the following:

1. The digital impersonation depicts the person in a sexual act or shows the unclothed breasts, buttocks or genitals of the person, and appeals to prurient interests, and offends local community standards
2. The plaintiff "was not a public figure, including a limited public figure, at the time the cause of action accrued"
3. The publication was made with actual knowledge that the content was a digital impersonation, or the publisher failed to take reasonable corrective action within 21 days after gaining actual knowledge

Subsection K sets the standard of proof for injunctive relief and damages at clear and convincing evidence. The public-figure bar in element 2 is the one that decides most cases: a candidate on the ballot is a public figure, so the political candidates the law is best known for protecting can generally obtain declaratory relief only, not damages.

Limitations and criticism:

Legal analysts have noted a weakness in the law's structure: a publisher who did not know a deepfake was fake at the time of publication has 21 days after gaining actual knowledge to take "reasonable corrective action," which the statute defines as removing or disabling access to the impersonation and publishing a correction to the same audience. A damaging deepfake can therefore circulate for nearly three weeks before that publisher faces exposure to damages.

Two limits on that criticism are worth stating plainly, because the 21-day window is often described as a blanket escape hatch and it is not one. First, element (I)(3) is written in the disjunctive: it is satisfied either by actual knowledge at the time of publication or by a failure to correct within 21 days. A publisher who knowingly publishes a digital impersonation meets the element through the first branch, and a later takedown does not cure it. Second, the corrective-action language appears only in subsection I, so it does nothing to the declaratory-relief track under subsections A through F, which remains available either way.

SB 1359: Candidate Deepfake Disclosure (A.R.S. § 16-1024)

Arizona passed a second deepfake law in 2024 that is frequently overlooked. Senate Bill 1359 (Laws 2024, Chapter 199) was approved by the Governor on May 29, 2024 and is codified at A.R.S. § 16-1024. It is Arizona's only affirmative AI-labeling requirement, and it operates independently of the § 16-1023 cause of action. Both 2024 deepfake bills were drafted to add a section numbered 16-1023, and the revisor renumbered SB 1359's provision to § 16-1024.

What it requires:

Within 90 days before an election at which a candidate for elected office will appear on the ballot, a "creator" may not create and distribute a synthetic media message the creator knows is a "deceptive and fraudulent deepfake" of that candidate unless the message includes "a clear and conspicuous disclosure that conveys to a reasonable person that the media includes content generated by artificial intelligence."

Key definitions:

• A creator is any person that uses artificial intelligence or other digital technology to generate synthetic media. It does not include the provider of the technology used to create it.
• A deceptive and fraudulent deepfake is synthetic media that depicts a candidate with knowledge of its falsity and intent to injure the candidate's reputation, and that is calculated to mislead a reasonable person into concluding a real individual said or did something they did not.

Exemptions:

The section does not apply to media that constitutes satire or parody, or to an interactive computer service as defined in 47 U.S.C. § 230.

Penalty:

A person who fails to make the required disclosure is liable for the per-day civil penalty prescribed by A.R.S. § 16-937(B), assessed for each day the deepfake is distributed without the disclosure.

Note the different triggers: § 16-1024 turns on a 90-day pre-election window and is satisfied by labeling the content, while § 16-1023 has no labeling option and uses a 180-day window as one route to preliminary declaratory relief.

SB 1295: AI Voice Fraud (Class 5 Felony)

Arizona's Senate Bill 1295 updates the state's criminal impersonation statutes to address AI-generated voice recordings, images, and videos used with intent to defraud. Under A.R.S. 13-2006(A)(4), "defraud" means making a false representation or material omission to deceive another person to gain a benefit.

Criminal penalties:

The Class 5 felony classification for AI-based impersonation carries more severe penalties than the existing Class 6 felony for traditional criminal impersonation, reflecting the legislature's view that AI tools amplify the potential harm of fraud-based impersonation crimes. The offense applies only where the AI-generated voice, image, or video was used with intent to defraud; it does not create a separate harassment-based offense.

Protected speech exemptions:

The law explicitly protects comedy, parody, art, criticism, and situations where a reasonable viewer can tell the media has been edited. This ensures that obvious satire and creative work are not treated as criminal fraud.

HB 2678: AI-Generated Child Exploitation Material

House Bill 2678, championed by Representative Julie Willoughby, was enacted in 2025 to close a critical loophole in Arizona's child exploitation laws. Before this bill, Arizona's criminal statutes did not explicitly cover AI-generated or digitally manipulated images depicting minors.

Key changes to existing law:

• Updates the legal definitions of "minor" and "visual depiction" to include computer-generated or digitally manipulated images that are indistinguishable from actual children
• Classifies AI-generated exploitation imagery as a dangerous crime against children when the depicted victim appears to be under age 15
• Applies existing child exploitation penalties to AI-generated content that meets the indistinguishability standard

The bill passed both chambers with strong bipartisan support, reflecting broad agreement that existing child protection laws needed updating for the AI era.

HB 2175: Healthcare AI Claim Denial Prohibition

House Bill 2175 makes Arizona among the first states in the nation to prohibit health insurers from using AI as the final decision-maker for medical claim denials. Governor Hobbs signed the bill into law, and the requirement took effect on July 1, 2026, codified at A.R.S. § 20-3103.

Core requirements:

• A medical director must individually review both claim denials and direct denials of prior authorizations involving medical necessity
• The medical director must exercise independent medical judgment and cannot rely solely on AI-generated recommendations
• The medical director must hold an active, unrestricted license to practice medicine in Arizona
• Written denials must include an explanation of why treatment was denied

The statutory text itself does not use the words "artificial intelligence." A.R.S. § 20-3103 reads in full: "Before a health care insurer may deny a claim that was submitted by a provider on the basis of medical necessity, the medical director shall individually review the denial. During each individual review, the medical director shall exercise independent medical judgment and may not rely solely on recommendations from any other source." The constraint on AI is the practical effect of that human-review mandate rather than an express ban on the technology.

What the law does NOT do:

HB 2175 does not ban AI in healthcare entirely. Insurers can still use AI tools to assist in processing claims and prior authorizations. The law specifically targets the final decision-making step, requiring that a qualified human physician makes the ultimate call on denials.

The Arizona Medical Association celebrated the law as a critical safeguard against the growing trend of insurers using algorithms to deny care without adequate physician oversight.

Arizona's Statewide Generative AI Policy

Beyond legislation, Arizona has taken administrative action to govern AI use within state government. The Arizona Department of Administration published its Generative AI Policy (P2000) on March 15, 2024, with updates enacted on October 24, 2024.

Guiding Principles

The policy establishes four core principles for state AI use:

1. Empowerment: Supporting the workforce to deliver services efficiently, safely, and equitably
2. Transparency and Accountability: Building trust and enabling collective learning
3. Fairness: Ensuring equitable outcomes from AI-assisted processes
4. Privacy and Security: Responsible experimentation that maintains control and respects individual privacy

Requirements for State Agencies

State agencies must comply with statewide IT and cybersecurity policies when using generative AI. Agencies are required to provide full attribution for which GenAI tools are used in their operations, ensuring public transparency about government AI use.

AI Steering Committee

Governor Hobbs also created an AI Steering Committee to guide future AI deployment across state government. The committee explores beneficial applications of AI while developing policies that promote transparency and ethical use.

Employee Training

The State of Arizona implemented employee GenAI training programs to ensure state workers understand how to use AI tools responsibly and within policy guidelines.

2026 Legislative Session: Pending AI Bills

Arizona's 2026 legislative session has introduced additional AI bills, though lawmakers have warned that the state's part-time legislature struggles to keep pace with the speed of AI development.

HB 2311: AI Chatbot Safety for Minors

House Bill 2311 passed the Arizona House by a vote of 43-13 on February 24, 2026, and received its second reading in the Senate on March 9. The bill would set new rules for how public AI chatbots interact with children and teenagers.

Key provisions:

• Disclosure requirements: Chatbot operators must inform minor users that they are interacting with AI, not a human, at the beginning of each session and at least once every three hours of continuous interaction
• Sexual content prohibition: Operators must implement "reasonable measures" to prevent AI from producing sexually explicit content, statements encouraging sexual conduct, or content that sexually objectifies the minor
• Engagement restrictions: Operators cannot offer points or rewards designed to encourage increased engagement by minor users
• Crisis response protocols: Chatbots must have plans for responding to users who discuss suicide or self-harm, including directing them to crisis hotlines
• Therapy limitations: AI chatbots are forbidden from claiming to be professional therapists

Enforcement: Only the Arizona Attorney General could enforce the law through civil penalties and court orders.

HB 2409: AI Education Program

House Bill 2409 would create a voluntary, statewide AI education program administered by the Arizona Department of Education. The program would offer summer classes open to all Arizonans, teaching students how to navigate the digital world safely and critically evaluate AI-generated content.

Attorney General Enforcement: The Grok Investigation

In January 2026, Arizona Attorney General Kris Mayes announced an investigation into xAI's Grok chatbot following reports that the platform generated sexually explicit images of minors. This investigation demonstrates Arizona's willingness to use existing legal authority to address AI harms.

Background

Reports emerged in late December 2025 and early January 2026 that Grok users could generate sexually explicit AI images depicting minors. Research from the Center for Countering Digital Hate estimated that Grok generated approximately 3 million sexualized images between December 29, 2025, and January 8, 2026, with roughly 23,000 allegedly depicting children.

Legal Basis

The investigation draws on Arizona's existing consumer protection and child exploitation laws, as well as the recently enacted HB 2678 which explicitly criminalizes AI-generated child sexual abuse imagery. Attorney General Mayes called on Arizonans who believe they were victimized by Grok to contact her office.

Broader Implications

Arizona's investigation is part of a multi-state effort, with California also launching a parallel investigation. This enforcement action signals that states are not waiting for federal guidance before acting against AI platforms that generate harmful content.

AI and Employment in Arizona

Arizona has not enacted specific legislation governing AI in employment decisions such as automated hiring tools. Unlike jurisdictions such as New York City (Local Law 144) or Illinois, which have passed targeted AI hiring regulations, Arizona relies on its existing employment and anti-discrimination framework.

However, employers using AI in hiring, scheduling, or performance evaluation should be aware that Arizona's anti-discrimination laws apply to AI-assisted decisions. Systems that produce biased outcomes could expose employers to liability under both state and federal civil rights laws.

Federal AI Policy and Arizona

Executive Order 14365

President Trump's Executive Order 14365 (December 2025) creates potential tension with Arizona's active AI regulatory approach. The order directs the DOJ to challenge state AI laws and threatens to withhold broadband funding from states with certain AI regulations.

Impact on Arizona's Laws

Arizona's enacted AI laws address areas that are largely protected under EO 14365's carve-outs:

• Child safety laws (HB 2678, HB 2311): Explicitly preserved under the executive order's child safety exception
• State government AI policy (P2000): Falls within the state procurement carve-out
• Healthcare AI regulation (HB 2175): May face more scrutiny, as healthcare regulation could be viewed as constraining commercial AI development
• Election deepfakes (HB 2394, SB 1359): The civil enforcement framework is less likely to trigger federal preemption concerns than criminal prohibitions

Arizona's targeted, sector-specific approach to AI regulation may prove more resilient to federal preemption challenges than comprehensive AI frameworks like Colorado's AI Act, because each Arizona law addresses a specific harm rather than regulating AI broadly.

Looking Ahead: Arizona's AI Regulatory Future

Arizona's trajectory in AI regulation is shaped by several factors:

Active enforcement: The Attorney General's investigation into Grok signals that Arizona will use existing legal tools to address AI harms, even as new legislation develops.

Legislative constraints: Arizona's part-time legislature faces challenges keeping pace with AI developments. Lawmakers have acknowledged this limitation, suggesting that regulatory approaches need to be flexible enough to adapt to rapidly evolving technology.

Healthcare leadership: With HB 2175 in force since July 1, 2026, Arizona is an early testing ground for healthcare AI regulation. The law's impact on insurance claim processing could influence similar efforts in other states.

Federal dynamics: Arizona's strong relationship with the federal government on border security and defense technology (including AI applications at military installations) creates additional considerations for how the state navigates federal AI policy.

More Arizona Laws

Explore other Arizona law topics on Recording Law:

• Arizona Recording Laws
• Arizona Data Privacy Laws
• Arizona Whistleblower Laws
• Arizona Background Check Laws
• Arizona Biometric Privacy Laws