The FTC warned that AI is a marketing term it will test

Start here

A company can call almost anything AI-powered. The US Federal Trade Commission's business guidance treats that label the way it treats any other advertising claim: something that has to be true and provable, not a vibe. Reading its own posts is a fast way to see what a marketer, or a curious consumer, should be asking about an AI claim.

What the documents say

The FTC's post 'Keep your AI claims in check', published 27 February 2023 by an attorney in its Division of Advertising Practices, states that AI is 'a marketing term,' and lists the questions the agency says it will ask: are you exaggerating what the product can do; are you claiming it works better than a non-AI product without adequate proof; are you aware of its foreseeable risks; and does the product actually use AI at all, since 'merely using an AI tool in the development process is not the same as a product having AI in it.' Three weeks later, a companion post, 'Chatbots, deepfakes, and voice clones: AI deception for sale', published 20 March 2023, addresses a different problem: tools used to deceive, rather than tools that are oversold. It states that 'the FTC Act's prohibition on deceptive or unfair conduct can apply if you make, sell, or use a tool that is effectively designed to deceive, even if that's not its intended or sole purpose,' and asks builders whether they should be making the product at all, whether they are mitigating foreseeable misuse before release, and whether they rely too heavily on after-the-fact detection of fakes.

Check this

The two posts together give a usable filter for any AI product claim: separate a capability claim, which needs evidence, from a deception risk, which needs mitigation regardless of intent. A reader evaluating a chatbot, an image generator or a voice tool can ask both questions directly: what evidence supports the performance claim, and what happens when someone inevitably uses this tool for exactly the harmful purpose its own design makes obvious.

What holds and what fails

What holds is the principle that existing consumer protection law applies to AI products without needing AI-specific statutes; both posts apply the ordinary FTC Act standard for deceptive and unfair practices to this newer category of claim. What these posts do not do is create new legal rules or name companies; they describe how the agency will apply existing law, and enforcement outcomes in any case depend on facts the posts do not cover. Readers should treat them as a statement of enforcement posture, not case law.

• Ask what evidence backs a specific AI performance claim before trusting it.
• Ask who could misuse a chatbot or generative tool, and what stops them, before you rely on it.
• Remember that using an AI tool to build a product is not the same as the product having AI in it.

Marketing hype around AI is not new, and neither is the agency's response: proving a claim is still the seller's job, and a tool's foreseeable misuse is still the seller's problem, whatever the technology is called.