Making Honest Claims About AI: What the FTC Expects

How FTC advertising rules apply to AI claims, what recent FTC actions targeted, and a checklist for businesses, agencies and resellers.

SuccessHD · Last reviewed October 23, 2026 · 6 min read

"AI-powered" has become one of the most common phrases in marketing. For a small business, an agency or a software reseller, it can be tempting to lean on it: the new phone assistant "never misses a lead," the scheduling tool "replaces your front desk," the marketing service "uses AI to find buyers."

The Federal Trade Commission (FTC) has made clear that claims about AI are held to the same standards as any other advertising claim. The agency's leadership and priorities have shifted since 2024, but the core rules have not: say what is true, have evidence before you say it, and do not mislead.

This guide explains those rules, summarizes recent FTC actions involving AI claims, and offers a checklist you can use before publishing. It is general information, not legal advice. For specific claims, talk with a qualified attorney.

The rules that already apply

In its Advertising FAQs for small businesses, the FTC summarizes three standards under the FTC Act: "Advertising must be truthful and non-deceptive," "Advertisers must have evidence to back up their claims," and "Advertisements cannot be unfair."

An ad is deceptive, the FTC explains, if it contains "a statement - or omits information - that is likely to mislead consumers acting reasonably under the circumstances" and the information is material, meaning it matters to the buying decision.

Two details trip people up:

Implied claims count. The FTC distinguishes express claims, stated directly, from implied claims that customers would reasonably take away. "Our AI receptionist works like a full-time employee" implies things about accuracy, availability and judgment that you would need to support.
Evidence comes first. The FTC's Policy Statement Regarding Advertising Substantiation, issued in 1984, says "advertisers and ad agencies" must "have a reasonable basis for advertising claims before they are disseminated." Gathering proof after a complaint arrives is too late.

What FTC enforcement on AI claims has looked like

A few cases show the pattern.

Operation AI Comply (September 2024). The FTC announced a sweep of actions against companies using AI claims to deceive. Then-Chair Lina M. Khan said, "There is no AI exemption from the laws on the books." Most of the cases involved business-opportunity schemes promising earnings from AI-powered online stores.
An online legal-help service (final order January 2025). The FTC alleged the company's "robot lawyer" service "failed to live up to its lofty claims that the service could substitute for the expertise of a human lawyer." The order required a $193,000 payment, notice to past subscribers, and barred unsupported claims that the service can substitute for a professional.
A maker of an AI content detector (final order August 2025). The company advertised the detector's accuracy. The order requires it to stop making accuracy or efficacy claims unless it has "competent and reliable evidence showing those products are as accurate as claimed."
A seller of "conversational AI" business packages (sued August 2025, settled March 2026). The FTC said the company claimed its product could "replace human customer service representatives," and alleged false earnings claims and false refund guarantees. The settlement bans the company and its owners from selling or marketing any business opportunity.
A media company and two other firms (final orders August 2026). The FTC alleged they falsely claimed to use a special algorithm that listened to conversations from smart devices to target ads; in fact, the FTC said, "the marketing service wasn't based on voice data, and consumers hadn't opted into this service." The firms agreed to pay a total of $930,000.

The direction has shifted, not the basics

Enforcement priorities do change. In December 2025 the FTC reopened and set aside a 2024 order against an AI writing-tool company, concluding that the complaint "failed to satisfy the legal requirements of the FTC Act" and that the order "unduly burdens artificial intelligence (AI) innovation." The director of the FTC's Bureau of Consumer Protection said the agency is focused on "targeting fraud and tangible consumer harm."

For businesses, the takeaway is not that AI claims are now unregulated. The AI-related cases the FTC has brought or finalized since 2025, including the content-detector, business-opportunity and smart-device advertising matters, involve the same basics: claims about what a product can do, what it will earn you, or how it handles consumer data, made without the evidence to back them up.

Claims that deserve a second look

Before publishing, check these common types of AI claims:

Capability claims: "replaces your receptionist," "handles every call." Can you show it does, under real conditions, for typical customers?
Performance numbers: accuracy rates, answer rates, response times. Where did the number come from, under what conditions, and is it current?
Results and earnings: more leads, more revenue, time saved. Results claims need support that matches what a typical customer can expect, not your best case.
"AI-powered" itself: Does the product actually use the technology in the way the ad suggests? In the smart-device advertising matter above, the FTC said "the marketing service wasn't based on voice data."
Human versus machine: If customers interact with an automated assistant, do not suggest they are talking to a person.
Privacy and consent: Claims that customers "opted in" or that data is "never shared" must match what actually happens.

A note for agencies and resellers

If you sell or white-label someone else's AI product, the claims in your marketing are still your claims. The substantiation policy names both advertisers and ad agencies. Practical steps:

Ask the provider for the testing behind any number you plan to repeat, in writing.
Do not round up or generalize vendor figures. A result from one industry may not hold in another.
Keep a file of the evidence for each claim, with dates, and review it when the product changes.
Train your sales team on what they can and cannot promise.

A pre-publication checklist

Is every factual statement true today, for this product as sold?
What would a reasonable customer take away, including anything implied?
Do we have the evidence now, and does it match the strength of the claim?
Have we disclosed limits a customer would want to know, such as when a person takes over?
Would we be comfortable showing the evidence file to a regulator or a customer?

Key takeaways

AI claims follow the same FTC rules as any ad: truthful, not misleading, and supported by evidence.
Substantiation must exist before the claim is made, and implied claims count.
Recent FTC actions targeted capability, accuracy, earnings and consumer-consent claims involving AI.
Enforcement priorities shifted in 2025, but cases against unsupported AI claims have continued.
Agencies and resellers are responsible for the claims in their own marketing.
This is general information, not legal advice.

Making Honest Claims About AI: What the FTC Expects

The rules that already apply

What FTC enforcement on AI claims has looked like

The direction has shifted, not the basics

Claims that deserve a second look

A note for agencies and resellers

A pre-publication checklist

Key takeaways

Sources

More on AI Agents & Automation

AI Receptionists and Voice Agents: What They Can and Can't Do

Human in the Loop: Keeping People in Charge of AI Agents

Small Business Workflow Automation: Triggers, Actions and Guardrails