The FTC and AI-Generated Reviews: What the Endorsement Rules Mean for Writers

Since October 21, 2024, faking consumer reviews or testimonials — including generating them with AI — has been a violation of a federal rule carrying civil penalties of up to $51,744 each; what remains legal is using AI to help write about experiences that are genuinely real, by people who genuinely exist, with any paid or insider relationship clearly disclosed. If you write reviews, testimonials, or roundups for money, that one sentence is most of what you need. The rest is knowing where the lines sit and what happens to people who cross them.

Key takeaways

• The FTC’s Rule on the Use of Consumer Reviews and Testimonials took effect October 21, 2024 and explicitly covers AI-generated fakes — reviews from nonexistent people or people with no real experience of the product.
• Penalties run up to $51,744 per violation, and in a fake-review operation each review can be a violation.
• The 2023-updated Endorsement Guides still apply regardless of authorship: material connections must be disclosed, and endorsements must reflect honest, actual experience.
• The FTC has pursued AI tools themselves — its Operation AI Comply sweep included Rytr, whose generator could mass-produce detailed testimonials from a keyword.
• For writers, the compliance test is simple: real experience, real person, real disclosure. AI assistance with the wording of a truthful review is not the target.

What changed in 2024, and why AI is named

Fake reviews were already deceptive practices the FTC could chase case by case. What the agency lacked was a rule that made penalties automatic and predictable. In August 2024 it finalized the Rule on the Use of Consumer Reviews and Testimonials, and when it took effect that October, the economics of review fraud changed: violations of a trade rule carry civil penalties, currently up to $51,744 each, without the FTC having to build a full deception case from scratch every time.

The rule’s core prohibitions, in plain terms:

• Fake reviews and testimonials. Writing, buying, or selling consumer reviews that misrepresent the reviewer’s existence or experience. The rule text specifically contemplates AI-generated reviews — a testimonial from a person who was never a customer, or never a person at all.
• Undisclosed insider reviews. Officers, managers, employees, and their relatives reviewing their own company’s products without disclosing the relationship.
• Fake independence. A company running a review website that claims to be independent while covering its own products.
• Review suppression. Using threats or intimidation to remove negative reviews, or presenting a curated subset as if it were all the reviews.
• Fake social proof. Buying or selling phony followers, views, and similar indicators for commercial purposes.

AI is named because it collapsed the cost of the first category. A review farm used to need people; now a single operator can generate ten thousand distinct, plausible, five-star experiences overnight. The FTC didn’t ban the tool — it made the output model-agnostic: a fake review is a fake review whether a human or a model typed it.

The FTC has already gone after the supply side

If this reads as theoretical, the enforcement record says otherwise. In September 2024 the FTC announced Operation AI Comply, a sweep against deceptive AI claims and schemes. One target was Rytr, an AI writing assistant whose “testimonial and review” feature could take a keyword and emit detailed, specific consumer reviews — details the user never provided and no one had experienced. The FTC’s complaint alleged the service furnished users with the means to create false reviews; the settlement barred Rytr from offering that feature.

That case matters for writers in two ways. First, it shows the FTC will pursue the *means* of fabrication, not just the businesses posting fakes. Second, it draws the conceptual line clearly: the problem was never “AI wrote sentences.” It was “the substance was invented.” A model summarizing your real testing notes and a model inventing a customer are the same technology doing legally opposite things.

The Endorsement Guides didn’t go anywhere

Alongside the new rule, the older framework still governs everything endorsement-shaped. The FTC’s Endorsement Guides — updated in 2023 to address social media and, yes, fabricated reviews — rest on two requirements that authorship can’t dodge:

1. Disclose material connections. Payment, free products, employment, family ties. If a brand sponsored your post, that disclosure is required whether you wrote it, your assistant wrote it, or a model drafted it. “The AI forgot the #ad” is not a defense; the obligation belongs to the humans and companies involved.
2. Honest opinion, actual experience. An endorsement must reflect what the endorser actually believes and, where it claims use, actual use. AI-generated enthusiasm about a product you’ve never touched fails this test in the exact way a human-ghostwritten fake would.

For working writers, this connects to a commercial truth we’ve covered before: affiliate reviews built on firsthand experience aren’t just legally safer — they’re the only kind search engines and readers still reward.

A working compliance checklist for writers

If review-adjacent content is part of your income, run each piece through five questions:

1. Did someone real have this experience? You, or a documented customer whose words you’re presenting accurately. Keep the evidence — order confirmations, photos, interview notes, test logs.
2. Does the byline match reality? No invented reviewers, no composite “customers,” no attributing AI-polished words to a person who never approved them.
3. Are material connections disclosed? Clearly, near the claim, in language a normal reader understands. Not buried on a disclosures page.
4. Would the subject of every factual claim survive a challenge? AI drafts hallucinate specifics — features, results, timelines. Verify anything checkable before your name goes on it.
5. Is anyone asking you to fake it? A client requesting invented testimonials is asking you to participate in a rule violation with per-violation penalties. Decline in writing.

Within those lines, AI is a drafting tool like any other. Whether to *say* you used one is a separate, subtler question — we’ve written about whether disclosing AI is a competitive disadvantage — but the FTC’s concern is fabrication, not assistance. It’s also worth knowing how your copy reads to the screening tools brands increasingly use: running a draft through a free AI detector shows you what an editor’s compliance check will see, and PaperBleach plans cover teams doing it at volume. There’s more writing on AI ethics and policy on the blog.

Frequently Asked Questions

Is it illegal to use AI to write a product review?

Using AI to help word a review of a product you genuinely used is not what the FTC targets. The 2024 rule bans reviews that misrepresent the reviewer’s experience or existence — a review from someone who never used the product, a testimonial from a person who doesn’t exist, an insider review hiding the relationship. AI is called out because it industrializes those fakes, not because the tool itself is banned. The legal question is whether the experience and the reviewer are real, not whether a model helped with the sentences.

What exactly does the FTC’s 2024 fake review rule prohibit?

The Rule on the Use of Consumer Reviews and Testimonials, effective October 21, 2024, bans writing, selling, or buying fake consumer reviews and testimonials, including AI-generated ones attributed to nonexistent people or people with no real experience. It also bans undisclosed insider reviews from officers, employees, and their relatives, company-run review sites posing as independent, suppressing negative reviews through threats or deceptive gating, and buying fake social media indicators like followers or views. Each violation can draw civil penalties.

What are the penalties for fake AI-generated reviews?

The rule allows the FTC to seek civil penalties of up to $51,744 per violation, adjusted over time for inflation — and in a review-farm scenario, each fake review can count. The FTC has also shown it will pursue the tools themselves: in its September 2024 ‘Operation AI Comply’ sweep, it took action against Rytr, an AI writing service whose review generator could produce detailed testimonials from a keyword, alleging it provided the means to create false reviews. The commercial lesson is that both buyers and suppliers of fake reviews are exposed.

Do the FTC Endorsement Guides apply if AI wrote the post?

Yes, fully. The Endorsement Guides, updated in 2023, are about relationships and truthfulness, not authorship. If a brand paid you, gave you free product, or employs you, that material connection must be clearly disclosed whether you wrote the post at your desk or generated it with a model. An endorsement must also reflect the endorser’s honest opinion and actual experience. AI changes none of that — it just makes it easier to produce copy that violates it at scale, which is exactly why the FTC modernized its rules.

How can freelance writers use AI on review content safely?

Keep three receipts. First, real experience: only write reviews and testimonials grounded in genuine use — yours or a documented customer’s — and keep the evidence, like photos, order confirmations, or interview notes. Second, real identity: never attribute words to a person who doesn’t exist or didn’t approve them. Third, real disclosure: state material connections plainly and honor client obligations to do the same. Within those lines, using AI to structure or polish is a drafting choice, not a violation. If a client asks you to invent reviews, that request is now asking you to break a federal rule.

The bottom line

The FTC didn’t outlaw AI writing — it outlawed the lie AI made cheap. Reviews from people who exist, about experiences that happened, with relationships disclosed: all of that remains legal however you draft it. What ended in October 2024 is the era when inventing the rest was merely tacky. Now it’s priced, per violation, and the agency has shown it will collect.