FTC AI Enforcement Actions 2026: Real Cases, What Gets Fined, and Your Checklist

Image: Pexels, used under the Pexels License.

TL;DR: The FTC has brought at least 12 AI-related cases since December 2023 under Section 5 of the FTC Act, with no AI-specific statute. The table below lists each one with its date, amount, and FTC source. The clearest pattern is claims the company could not back up: DoNotPay's "robot lawyer" ($193,000), Workado's "98 percent" AI detector (53 percent on general content in testing), IntelliVision's "zero bias" claim, and Cox Media's "Active Listening" ($930,000 across three firms in May 2026). One case was reversed: the FTC set aside the Rytr order on December 22, 2025, citing America's AI Action Plan. Use the 7-point checklist at the end to check your own AI claims.

Updated October 8, 2026. We re-checked every case on this page against FTC press releases. The earlier version named "Automaters AI" as an Operation AI Comply case (the fifth case was FBA Machine), dated the AI Action Plan to February 2025 (it was released July 23, 2025), described the Sitejabber order as an AI-disclosure rule (it was about ratings collected before customers received their orders), and cited an FTC guidance document called "AI and Your Business" that we could not find. A state attorney general section with unsourced counts has been replaced with two sourced state cases.

The Federal Trade Commission was created on September 26, 1914, and Section 5 of the FTC Act still does the work in AI cases: "unfair or deceptive acts or practices in or affecting commerce, are hereby declared unlawful." The FTC has not needed a new AI law to bring the cases below.

FTC AI cases, December 2023 to May 2026

Operation AI Comply, announced September 25, 2024, covered DoNotPay, Ascend Ecom, Ecommerce Empire Builders, Rytr, and FBA Machine. The FBA Machine action itself was taken in June 2024 and folded into the announcement.

For a broader list that also covers the EU, the EEOC, and state attorneys general, see the AI Enforcement Tracker.

The legal theory: Section 5 deception and unfairness

Section 5 has two consumer-protection prongs.

Deception. Under the FTC's 1983 Policy Statement on Deception, the Commission "will find deception if there is a representation, omission or practice that is likely to mislead the consumer acting reasonably in the circumstances, to the consumer's detriment." Most cases in the table are deception cases about what an AI product could do: replace a lawyer, detect all weapons, perform with zero bias, detect AI text at 98 percent, or hear conversations through smart devices.

Unfairness. Section 5(n) says the FTC may not call a practice unfair "unless the act or practice causes or is likely to cause substantial injury to consumers which is not reasonably avoidable by consumers themselves and not outweighed by countervailing benefits to consumers or to competition." Rite Aid is the clearest AI example: the FTC said the retailer deployed facial recognition without reasonable safeguards, and its consumer protection director called the order a warning against "unfair biometric surveillance and unfair data security practices."

What the FTC orders actually require

Image: Pexels, used under the Pexels License.

The earlier version of this page described an FTC guidance document called "AI and Your Business". We could not find it on ftc.gov, and an older FTC business blog post on AI claims we tried to cite now returns "Page not found". The orders themselves are a better guide, because they are what a company actually signs:

• Evidence before the claim. Workado may not make effectiveness claims about its detection products unless it "maintains competent and reliable evidence showing those products are as accurate as claimed." IntelliVision may not make claims about the effectiveness, accuracy, or lack of bias of its facial recognition unless it has competent and reliable testing. DoNotPay may not claim it can substitute for any professional service "without evidence to back it up."
• Notice to customers. DoNotPay must notify 2021 to 2023 subscribers about the limits of its law-related features. Evolv must let certain K-12 school customers cancel their contracts.
• Money where consumers paid. DoNotPay's $193,000, accessiBe's $1 million, and Cox Media's $930,000 are all tied to settlements.
• Reports for years. Workado must file compliance reports one year after the order and every year for the following three years.

The pattern is simple. If your marketing gives a number (98 percent, zero bias, all weapons), the FTC will ask for the test behind it.

Fake reviews: the rule with civil penalties

On August 14, 2024, the FTC announced a final rule banning fake reviews and testimonials. It "addresses reviews and testimonials that misrepresent that they are by someone who does not exist, such as AI-generated fake reviews," and it took effect on October 21, 2024. The rule "authorizes courts to impose civil penalties for knowing violations." The FTC's maximum civil penalty for these violations is $53,088 per violation at 2025 levels, and the FTC said those amounts "will remain unchanged during 2026."

The Rytr reversal does not change the rule. It set aside one order against a tool maker; the rule still applies to businesses that create, buy, or spread fake reviews.

Algorithmic pricing: statements and studies, not AI cases

On March 28, 2024, the FTC and the Justice Department filed a statement of interest in a hotel room pricing case, explaining "that hotel companies cannot use algorithms to evade antitrust laws." In July 2024, the FTC sent 6(b) orders to intermediaries about surveillance pricing, and on January 17, 2025 it released interim staff findings. None of these is an enforcement order against a company for AI pricing, and none appears in the table above.

The 2025 policy shift and what survived it

The White House released "Winning the AI Race: America's AI Action Plan" on July 23, 2025. On December 22, 2025, the FTC reopened and set aside the 2024 Rytr order, saying the complaint "failed to satisfy the legal requirements of the FTC Act" and that the order "unduly burdens artificial intelligence (AI) innovation in violation of the Trump Administration's Artificial Intelligence Executive Order and America's AI Action Plan." Rytr consented.

The same release says the FTC "will continue to hold accountable actors that use AI to violate the law or deceive consumers about the capabilities of their generative AI." The record after the shift matches that line: the Workado final order (August 28, 2025) and the Cox Media settlement (May 21, 2026) are both deception cases about AI capability claims.

State attorneys general use their own laws

Image: Pexels, used under the Pexels License.

State attorneys general bring AI cases under state law without waiting for the FTC. Two examples:

• Texas, September 18, 2024. Attorney General Ken Paxton announced a settlement with Pieces Technologies, a Dallas healthcare AI company, resolving allegations that it made "false and misleading statements about the accuracy and safety of its products." At least four Texas hospitals had been sending patient data to Pieces.
• Massachusetts, July 10, 2025. Attorney General Andrea Joy Campbell announced a $2.5 million settlement with student lender Earnest Operations over lending practices, "including through the use of artificial intelligence (AI) models that could lead to disparate harm to Black, Hispanic, and non-citizen applicants and borrowers."

Neither case needed an FTC action first. If you sell AI to consumers or use it in lending, health care, or hiring, state consumer protection and fair lending laws are a second front.

Image: Pexels, used under the Pexels License.

7-point compliance checklist for companies using AI in consumer-facing products

Each item should have an owner who can show the current state and produce documents.

1. Inventory consumer-facing AI. For each AI feature: what it does, what data it uses, what it outputs, and which decisions it affects. You cannot check claims you have not listed.

2. Match every AI claim to a test. List each accuracy, bias, or capability claim in marketing, sales decks, and product pages. Attach the test that supports it. No test, no claim. This is the Workado and IntelliVision lesson.

3. Remove "replaces a professional" language. DoNotPay's order bars claims that its product substitutes for any professional service without evidence. Check chatbot copy for lawyer, doctor, or therapist framing.

4. Check review and testimonial sources. No AI-written reviews presented as customer reviews, no reviews collected before customers have the product (Sitejabber), and disclosed ties to reviewers (accessiBe).

5. Test AI that sorts people. For facial recognition, screening, lending, or hiring tools, keep records of accuracy across demographic groups. Rite Aid and Earnest are both about uneven harm to groups of people.

6. Offer a human path for high-stakes decisions. For credit, housing, employment, or access decisions, give people a way to reach a human and record how it works. This is not an FTC order term in the cases above; it is a practical safeguard.

7. Prepare for an FTC inquiry. Name the person who answers regulators, where the AI documents live, and who reviews them before they go out. Decide this before a letter arrives.

The AI regulation deadline calendar 2026 tracks related state deadlines, and the FTC AI marketing claims checklist goes deeper on item 2.

Our take

Based on the FTC releases cited here, the enforcement line through 2026 is narrower than the 2024 headlines suggested and steadier than the Rytr reversal implies. The FTC dropped a case against a tool that could be misused. It kept bringing cases against companies that said their AI could do something it could not. If your marketing makes a measurable AI claim, the most useful thing you can do this quarter is file the test that backs it up.

How we checked this

On October 8, 2026 we read each FTC press release or case page linked in the table, the Section 5 text at 15 U.S.C. 45, the FTC's 1983 Policy Statement on Deception, the August 2024 reviews rule release and the FTC's rule Q&A, the 2025 civil penalty release and the 2026 Federal Register notice, the White House AI Action Plan release, and the Texas and Massachusetts attorney general releases. We removed claims we could not source, including a count of state AI task forces, two Illinois facial recognition cases, an "AI and Your Business" guidance document, and a statement about company size. We looked for a Reddit discussion of Operation AI Comply to quote; Reddit blocked our reads on October 8, 2026, so this page has no community quotes.

Last reviewed: October 8, 2026.

Related Reading

• Treasury's first outbound investment fine: $200K against Amidi for a China AI deal

• SEC AI Examination 2026: What Examiners Will Ask Your Team About AI

• AI securities class actions 2026: disclosure checklist

• AI slowdown antitrust suit: 3 checks for your vendor contracts

• AI Hallucination Sanctions Tracker 2026: $145K in Fines, One Fix

• AI Cybersecurity Clearinghouse deadline hits today: what it means for your team

• AI Enforcement Tracker: every real AI regulatory action (FTC, EU, EEOC, state AGs), sourced

• Dario Amodei AI regulation proposals June 2026: binding rules, a jobs tax, and what your team should watch

• Pentagon used Grok AI to target 2,000 Iranian sites in 96 hours: Gillibrand bill and AI oversight policy

• German court rules Google liable for false AI Overviews: implications for AI deployers

• FTC AI marketing claims checklist 2026

• AI enforcement: multi-channel risk guide

• AI governance checklist 2026

• AI regulation deadline calendar 2026

• AI acceptable use policy template

• AI governance guide for small teams

• Maryland AI algorithmic pricing law 2026

• Tennessee ELVIS Act AI voice likeness compliance 2026

• Trump's June 2026 AI executive order: what compliance teams and federal contractors must do

• GPT-5.6 stagger: 5 AI procurement policy updates after the White House intervened

• FTC's $930K Active Listening settlement: what AI-washing enforcement now looks like

• FTC Says Your AI Vendor May Be Breaking the Law If It Secretly Steers Outputs

• RentGrow FTC Settlement: AI Tenant Screening FCRA Checklist 2026

• White House Now Gates Frontier AI Models: What EO 14409 Means

• Anthropic's $1.5B Copyright Settlement: Training Data Has a Price Now

• France Antitrust: OpenAI, Google, Anthropic Hold 84% of AI Agents

• Starbuck v. Google: AI Defamation Survives Dismissal, Heads to Trial

• FTC Says Colorado's AI Bias Law May Be Preempted. Comment by July 31.

• China's Moonshot Accused of AI IP Theft. Treasury Threatens Sanctions.

• Anthropic's Project Panama: Book Destruction, FTC Risk, and You

• OpenAI's $3.2M DOJ Settlement: PERM Violations and HR Vendor Risk

• Judge voids Pentagon Anthropic blacklist: AI vendor compliance guide

• Cleo AI FTC Refunds: $15.8M by PayPal, Dates and Who Gets Paid