In plain English
- What happened
- In September 2024, the Federal Trade Commission announced a sweep of enforcement actions it called Operation AI Comply, targeting companies for making exaggerated or unsupported claims about what their AI products could do, including one that had marketed itself as the world's first robot lawyer. The agency's 2023 business guidance on AI marketing had already warned that claims about artificial intelligence are judged by the same truth-in-advertising rules as any other product claim.
- Why it matters to your business
- A small business that calls a tool or service 'AI-powered' is making a factual claim, not a marketing flourish, and the FTC treats it that way. If the claim turns out to be exaggerated, the business that wrote the ad copy is the one on the hook, not the AI vendor whose model sits behind it.
- Do this
- Reread any marketing line that mentions AI and check whether it describes something the tool has actually been tested to do, using the FTC's 2023 guidance as a checklist.
No, not automatically, and the Federal Trade Commission has already acted on this. The agency's September 2024 sweep, Operation AI Comply, brought cases against several companies the FTC said had overstated what their AI products could actually do for customers.
What the FTC actually did
One of the companies named in the sweep had marketed itself as offering a 'robot lawyer' that could handle legal matters for consumers. The FTC's complaint said the company had not done enough to back up what that service could reliably deliver, and the resulting order required changes to how it could describe its capabilities going forward and notice to affected customers.
This was not a one-off. The FTC's February 2023 business guidance laid out, in plain terms, what the agency looks for before that kind of case ever gets filed: whether a business is exaggerating what its AI tool can do, making claims about accuracy or effectiveness it has not tested, or deploying AI without assessing the harm it could cause customers if it gets things wrong.
The FTC's own guidance says AI marketing claims are evaluated under the same truth-in-advertising standard as any other product claim — 'AI' is not a special category with a lower bar.
What counts as an AI claim
It is broader than most business owners expect. Calling a chatbot, a scheduling tool, or a marketing service 'AI-powered,' 'state-of-the-art,' or 'as accurate as a human expert' are all claims the FTC's guidance says a business should be able to support with evidence, not just confidence in the vendor's own marketing.
That matters because most small businesses are not writing their own AI models. They are reselling, wrapping, or recommending a tool someone else built, and the tendency is to repeat whatever the vendor's sales page says. The FTC's guidance is clear that borrowing a vendor's claim does not transfer the vendor's liability away from the business that repeats it to its own customers.
Reality check
This is not a reason to scrub the word 'AI' from every webpage. A business that genuinely uses AI to do something — answer questions faster, draft a first pass of a document, flag a scheduling conflict — can say so. The risk sits specifically with claims of accuracy, expertise, or outcomes nobody has actually measured.
What to check before the next claim goes out
- 1
List every place AI is mentioned in your marketing
Website, social posts, flyers, email signatures. A quick search of your own site for the word 'AI' usually turns up more mentions than expected.
- 2
Match each claim to something you can point to
If a page says a tool is 'accurate' or 'expert-level,' find the specific thing that backs that up — a test you ran, a result a customer saw — or soften the claim to what is actually true.
- 3
Separate what the vendor says from what you are promising
A vendor's marketing page is not proof for your own ad. If you are repeating a vendor's claim as your own, treat it as your claim to defend.
- 4
Note who checks AI-related copy before it publishes
The FTC's guidance applies at the point a claim reaches a customer, so the review that matters is the one before publication, not after a complaint arrives.
None of this requires a lawyer on retainer or a compliance department. It requires treating the word 'AI' in an ad the same way a business would treat the word 'guaranteed' or 'certified' — as a claim someone could ask to see proof of.
Sources
