← All retail signals
Governance

The Policed Review

The fake review faces two lines of defence: a US ban naming AI-generated reviews, with penalties up to $53,088 per violation, and detectors one study puts above 90% accuracy.

What is it?

The FTC’s rule banning fake and AI-generated reviews is now in force, with civil penalties reaching $53,088 per violation. Enforcement — warning letters, a live docket — is arriving years before independent measurement shows AI-generated reviews make up more than a small, single-digit share of any storefront.

Why now?

A 2025 warning wave followed the 2024 rule, and 2026 research finally measured the target: roughly three percent of top Amazon reviews carry AI-generation signatures, skewed toward five-star ratings and slipping past the ‘Verified Purchase’ badge meant to catch exactly this.

Who’s building it?

Federal regulators write and enforce the rule; platforms build their own detection layers around the same target.

The FTCAmazonDetection vendorsPlatform trust teamsWarned sellers
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Regulators — the FTC, enforcing a live rule with real penalties.
Platforms — Amazon and rival marketplaces, whose own review scores are now under scrutiny.
Sellers — the ten firms warned in December, and whoever a scan flags next.
What’s at stake
The $53,088-per-violation ceiling is real and verified, but it prices a problem regulators have not yet measured at scale.
Which platform or seller draws the next warning letter.
Whether enforcement is outpacing the fraud it targets.
What’s unsettled
Whether the 3 percent figure holds outside Amazon’s ten categories.
Whether warning letters ever turn into levied civil penalties.
Whether marketplaces disclose their own AI-review detection rates the way regulators now require of sellers.
02 · How the demand matureda progression, not a timeline
Rule
a final rule bans fake and AI-generated reviews outright, effective nationwide.
Warning
ten companies receive letters naming the statute before any fine lands.
Penalty
the next step is a company actually paying the $53,088-per-violation maximum, not a warning.
Precedent
the first paid penalty becomes the number every seller’s compliance budget plans around next.
03 · Where the market goesthe forks, not a prediction
→Whether the FTC actually levies a payable civil penalty against one of the ten warned companies, or the docket stays at warning letters indefinitely.
→Whether marketplaces publish their own AI-review detection rates instead of leaving that number to outside researchers.
→Whether the three percent aggregate figure holds inside beauty specifically, or beauty proves higher once it’s actually measured.
→Whether a state attorney general brings the first case a company can’t simply settle.
The rule and its penalty ceiling are real, arriving years ahead of the fraud they punish — and no company has paid that price yet.
04 · The Enforcement Docket deskone rule, ten warnings, zero fines yet
Where The Rule Started
A unanimous vote made this official.
5-0 FTC vote; rule effective 60 days after publication.
FTC
The Detection Layer Behind It
Academic detectors already spot fake reviews reliably.
93% accuracy on Amazon review data, 91% on Yelp.
Tech Xplore
The Same Fight, A Different Platform
Reddit built its own defence at real scale.
23 million spam views blocked daily, per Reddit’s own count.
TechCrunch
05 · The Wire
01
The FTC’s rule banning AI-generated reviews passed 5-0.
FTC ↗
02
The FTC warned 10 companies, citing penalties up to $53,088.
FTC ↗
03
A hybrid detection model reached 93% accuracy distinguishing genuine from deceptive reviews on Amazon data and 91% on Yelp.
Tech Xplore ↗
04
Reddit says it blocks 23 million spam views per day
TechCrunch ↗
The panel on this signal
They agreed on
–A binding federal rule now names AI-generated reviews specifically, with penalties running into the tens of thousands per violation.
–Ten companies already received warning letters in December.
–Independent measurement puts AI-flagged reviews at a small, single-digit share of listings.
The marketplace fraud investigator disagreed on
–A ‘Verified Purchase’ badge was never built to catch AI-written text, so fraud investigators expected this exact pattern, not a surprise.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.