This article summarises publicly available guidance from regulators and official sources. It is general educational information only and does not constitute legal or professional advice. Requirements vary by jurisdiction. Consult your regional authority or a qualified professional for advice specific to your situation.
The FTC isn't targeting businesses for using AI, it's targeting businesses for making claims about AI products that don't hold up. Under Section 5 of the FTC Act, a claim is unfair or deceptive if it's likely to mislead a reasonable consumer and affects a decision they'd make, and the FTC has shown it will enforce this against AI marketing specifically, not just wait for a complaint before acting.
In short: The FTC treats unverified AI performance claims as a real enforcement target, real cases exist, not hypothetical risk. A claim your AI feature performs "like a human expert" or is highly accurate needs to be actually true and tested, not aspirational marketing language. The FTC has also drawn a specific line between genuine AI hallucinations (a technical limitation) and deliberately steering an AI system's outputs to mislead, the second is the more serious concern.
The plain-English answer
Using AI in your product is fine. Claiming your AI does something it doesn't reliably do is the problem, and it's the same standard the FTC has always applied to advertising generally, just now actively enforced against AI-specific claims. If your copy says your tool performs at a certain level, was tested a certain way, or replaces a category of professional expertise, and none of that was actually verified before you wrote it, that's the exposure, regardless of how common that kind of claim is across your competitors.
Why this matters for your business
A 12-person SaaS startup writing launch copy for a new AI feature is under real pressure to sound as compelling as competitors, and AI marketing copy across the industry leans heavily on confident, sweeping claims. That competitive pressure is exactly how unverified claims end up shipping, not through any intent to deceive, but because the claim sounded reasonable and nobody checked it against what the feature actually does.
The FTC's enforcement pattern shows this isn't a large-company-only risk. Small, genuinely well-intentioned companies have been the subject of real enforcement actions for exactly this pattern, a confident claim that was never actually tested against the standard it implied.
What the FTC has actually done
The FTC's AI industry page and its Operation AI Comply crackdown, announced in September 2024, targeted companies making AI claims that didn't hold up, including tools claiming to generate fake reviews and services claiming AI capabilities they didn't actually have. A specific, well-documented case worth knowing: DoNotPay's "AI Lawyer" case, finalised in February 2025. The FTC's complaint found DoNotPay hadn't tested whether its "AI lawyer" performed to the level of a human lawyer, and hadn't hired or retained attorneys to verify the quality of its law-related output, despite marketing it as a substitute for legal expertise. The final order required $193,000 in monetary relief, notice to past subscribers, and prohibited the company from making performance claims it can't back up going forward.
More recently, the FTC has sharpened its focus on a related but distinct issue: in July 2026, it issued a proposed policy statement addressing what it calls the "suppression of accuracy" in AI systems, specifically distinguishing deliberate output steering to achieve an undisclosed goal from genuine AI hallucinations, inaccuracies from technological limitations, which the FTC says generally don't raise the same concern on their own. The distinction matters: an AI tool that's occasionally wrong isn't automatically an FTC problem, a business that markets it as never wrong, or that deliberately tunes outputs to mislead, is a different matter entirely.
What this looks like in practice
Picture the marketing lead drafting launch copy for the new AI feature, landing on the line "works just like having a human analyst on your team." It reads well, it's the kind of claim every competitor in the space seems to make, and nobody on the product team has actually benchmarked the feature's output against a real analyst's work.
Reviewing the FTC's actual enforcement pattern changes the approach: rather than drop the claim entirely, the team runs a real comparison, the feature's output against a human analyst's on the same task, and finds it holds up reasonably well on routine cases but falls short on complex ones. The rewritten copy describes what was actually tested and verified, still a strong claim, but one that can survive scrutiny rather than one that sounded good and was never checked.
What you can do about it
A practical review process before any AI feature claim ships:
- For any specific performance claim, verify it's actually true, tested against a real benchmark, not just plausible-sounding.
- Avoid comparing your AI feature's output to a human professional's unless you've genuinely benchmarked it against one.
- Don't claim your AI is never wrong or fully reliable, acknowledge real limitations rather than implying perfection.
- If your AI system's output can be tuned or adjusted, be transparent about how and why, undisclosed output steering is a specific concern the FTC has flagged.
- Keep a record of what was tested and how, this is exactly the kind of evidence that matters if a claim is ever questioned.
If your business also publishes AI-generated content that customers or the public see, see our guide on who's legally liable for AI-generated content for the related but distinct question of liability for what the AI itself produces, not just claims about the AI product.
Methodology (Real-World, Verified)
We test AI tools against real SMB workflows: the tasks a 20-person business actually uses AI for, not enterprise demos. Pricing is verified at the vendor's published rates, with local-currency conversions noted where relevant. Compliance notes reference the legislation and regulatory guidance relevant to each article's region. Every tool is judged on one question: could a business with no dedicated IT department actually pick this up and use it on Monday morning.
Related reading: our AI governance by region.
Try our free AI Compliance Checker to check whether your AI tools meet your compliance obligations.
Does this only apply to companies selling AI products, or also businesses using AI internally?
It applies most directly to claims made to consumers or customers, whether that's a company selling an AI product or a business marketing an AI-powered feature or service. Internal AI use with no external claim attached carries much lower FTC exposure under this specific framework.
Is it enough that our AI vendor claims their tool is accurate?
Not necessarily, if you're the one making the claim to your customers, verifying it is your responsibility, not just the vendor's. Passing along an unverified vendor claim as your own marketing statement doesn't shift the FTC exposure away from your business.
What's the difference between a genuine AI hallucination and something the FTC would penalise?
The FTC has drawn this line explicitly: an occasional inaccuracy from a technological limitation generally doesn't raise the same concern as deliberately steering outputs to achieve an undisclosed goal, or marketing a tool as never wrong when it demonstrably isn't. The concern is about deception, not about AI being imperfect.
How big does a company need to be before the FTC takes an enforcement action seriously?
Company size hasn't been the deciding factor in the FTC's actual cases, several targeted smaller companies making specific, unverified claims. The exposure comes from the claim itself, not from company scale.
The information in this article is general in nature. It reflects a summary of publicly available guidance and does not constitute legal, privacy, or professional advice. Your obligations will depend on your specific situation, jurisdiction, and business circumstances. Do not rely on this article as a substitute for qualified legal or professional advice.
Also publishing AI-generated content your customers or the public will see?
Check the liability picture