Florida AI Legislation (Deepfake and AI Disclosure Laws)

Summary

As of July 4, 2026, Florida AI laws mean a narrow three-layer compliance set: Fla. Stat. § 106.145 political generative-AI ad disclosures, Fla. Stat. § 836.13 altered-sexual-depiction and platform-takedown rules, and FDUTPA penalties for deceptive AI conduct; no comprehensive Florida AI Act is in force.

Quick answer: what are Florida AI laws in 2026?

As of July 4, 2026, Florida has no single comprehensive AI act. A Florida-facing AI register should separate the following controls instead:

The Florida Senate passed SB 482, the 2026 Artificial Intelligence Bill of Rights proposal, but the official bill history says it died in House messages on March 13, 2026. That proposal remains a watchlist signal, not an enacted Florida AI compliance duty. Source: Florida SB 482 bill history, retrieved 2026-07-04.

Florida AI laws compliance map by use case

Florida does not use one statewide "AI system" definition. A compliance screen should start with the use case and then attach the Florida source that can actually reach it:

Political AI advertising: HB 919 and § 106.145

HB 919 created Fla. Stat. § 106.145 for AI use in political advertising. The statute applies when political advertising, electioneering communication, or another political advertisement contains images, video, audio, graphics, or other digital content created in whole or in part with generative AI, appears to depict a real person performing an action that did not actually occur, and was created with intent to injure a candidate or deceive regarding a ballot issue.

The required control is a prominent disclosure stating that the ad was created in whole or in part with generative AI. The format changes by medium: printed communications use bold 12-point text, video must be readable throughout and occupy at least 4 percent of picture height, internet public communications must be viewable without user action, and audio disclosures must run at least 3 seconds. Failure to include the disclaimer is treated as a first-degree misdemeanor in addition to other election-law civil penalties. Source: Fla. Stat. § 106.145, retrieved 2026-07-04.

Altered sexual depictions: § 836.13 and Brooke's Law

Florida's deepfake law is now broader than a posting-only rule. Fla. Stat. § 836.13 covers willfully generating, soliciting, promoting, or possessing with intent to promote a nonconsensual altered sexual depiction of an identifiable person. Each prohibited act or transaction is a separate offense, and a disclaimer inside the altered depiction is not a defense. Source: Fla. Stat. § 836.13, retrieved 2026-07-04.

The private civil-remedy hook matters for exposure modeling. An aggrieved person may seek injunctive relief, attorney fees and costs, and monetary damages equal to $10,000 or actual damages, whichever is greater, for violations of the generation or promotion provisions. Covered platforms also need a notice-and-removal process by December 31, 2025 and must remove a validly reported altered sexual depiction, plus known identical copies, as soon as possible and no later than 48 hours after receiving the request. A platform's failure to reasonably comply is treated as an unfair or deceptive act under FDUTPA. Source: Fla. Stat. § 836.13(7)-(8), retrieved 2026-07-04.

FDUTPA and AI-related consumer deception

Florida's Deceptive and Unfair Trade Practices Act is not AI-specific, but it is the enforcement container that Florida already uses for deceptive or unfair trade practices. For AI systems, the practical risk is misrepresentation: synthetic reviews, bot identity confusion, misleading generated claims, impersonation, or failure to honor a required takedown workflow.

The fixed penalty cap depends on the affected population and claim path. Section 501.2075 allows civil penalties of not more than $10,000 per willful violation. Section 501.2077 raises the penalty to not more than $15,000 per violation for willful practices victimizing or directed at senior citizens, persons with disabilities, military servicemembers, or covered military family members. Source: Fla. Stat. § 501.2075 and Fla. Stat. § 501.2077, retrieved 2026-07-04.

2026 Artificial Intelligence Bill of Rights watch item

SB 482 would have created an Artificial Intelligence Bill of Rights in chapter 501, including companion-chatbot controls, bot disclosure, deidentified-data duties, name-image-likeness changes, and artificial-intelligence instructional-tool notice provisions. The enrolled Senate version also contained government-contracting restrictions for AI technology providers connected to foreign countries of concern.

The key compliance point is status: the Senate passed CS/SB 482 as amended on March 4, 2026, the bill moved to the House in messages on March 5, 2026, and the official Florida Senate history records "Died in Messages" on March 13, 2026. It should be tracked as a 2027-session watch item, not assigned as a binding Florida AI obligation. Source: Florida SB 482, retrieved 2026-07-04.

What Florida AI compliance actually requires

In 2026, an AI system deployed to Florida consumers should be screened by use case, not by a generic high-risk-AI label:

1. Political content - determine whether generated content depicts a real person doing something that did not occur and whether the ad was intended to injure a candidate or deceive voters. If yes, preserve the disclaimer evidence under § 106.145.
2. Sexual or intimate imagery - block generation, solicitation, promotion, and possession-with-intent workflows for nonconsensual altered sexual depictions; maintain consent and escalation records under § 836.13.
3. Platform hosting - maintain a clear notice-and-removal process, a 48-hour takedown SLA for valid requests, and a duplicate-copy search process for covered-platform obligations.
4. Commercial AI claims - substantiate generated claims, chatbot disclosures, reviews, endorsements, and impersonation controls under FDUTPA.
5. Governance evidence - map each Florida use case into an AI compliance framework, assign ownership through an AI governance operating model, and use NIST AI RMF or ISO/IEC 42001 as the voluntary control baseline.

Florida's narrow per-statute approach contrasts with Texas TRAIGA, Colorado's scheduled ADMT framework, the Utah AI Policy Act, and Virginia HB 2094, which was vetoed. Washington SB 5838 shows another non-comprehensive path: a state task force rather than direct private-sector AI duties.

Frequently asked questions

Does Florida have a comprehensive AI law in 2026?

No. Florida has binding AI-specific provisions for political advertising and altered sexual depictions, plus FDUTPA exposure for deceptive or unfair AI-related conduct. SB 482 would have created a broader Artificial Intelligence Bill of Rights, but it died in House messages on March 13, 2026.

What is the main Florida AI disclosure rule?

The clearest AI-specific disclosure rule is Fla. Stat. § 106.145. Political ads and electioneering communications using qualifying generative-AI content must include the statutory AI disclosure in the required medium-specific format.

What does Brooke's Law require for AI deepfakes?

Fla. Stat. § 836.13 prohibits generating, soliciting, promoting, or possessing with intent to promote nonconsensual altered sexual depictions of identifiable people. Covered platforms must maintain a notice-and-removal process and remove validly reported content, including known identical copies, within 48 hours.

What penalty number should a Florida AI register use?

Use a layered penalty model. Section 836.13 creates separate offenses and a civil action with at least $10,000 or actual damages for covered generation or promotion violations. FDUTPA adds up to $10,000 per willful violation, or up to $15,000 per violation for protected-population cases under § 501.2077. The Atlas penalty field uses $15,000 as the highest fixed per-violation civil penalty currently tied to this Florida record.

Should Florida AI systems use an impact assessment?

Florida's enacted AI-specific statutes do not impose a Colorado-style impact-assessment duty. An impact assessment can still be useful evidence for FDUTPA risk, platform moderation, political-ad review, and broader NIST or ISO governance, especially when the same AI system also operates in Colorado, Texas, California, NYC, Illinois, Utah, or Connecticut.