When a "robot lawyer"
couldn't practice law
DoNotPay called itself "the world's first robot lawyer" and claimed its AI could replace human attorneys. The FTC investigated and found no evidence it could practice law — or substantiate any of its marketing claims. The settlement set the first major precedent for regulating AI marketing.
The AI that marketed itself as a lawyer
DoNotPay built a chatbot for contesting parking tickets and small claims. Over time its marketing expanded: "the world's first robot lawyer," an AI that could "sue anyone," "beat any lawyer," and replace expensive attorneys for matters from divorce to immigration.
None of it was tested in court. The product was never evaluated against real legal-work benchmarks, no attorneys reviewed its outputs for accuracy, and there was no evidence the AI could perform the legal analysis it claimed. The "robot lawyer" was a marketing position, not a measured capability.
The FTC opened an investigation. It found DoNotPay could not substantiate any of its core claims — that the service was equivalent to a human lawyer, that it could replace one, or that it had been tested for legal competence. The agency characterized the campaign as deceptive.
In January 2024, DoNotPay agreed to a settlement: a $193,000 payment and a permanent ban on claiming its service can be used instead of a lawyer or that it can match the quality of a human attorney — unless it has competent and reliable evidence.
What DoNotPay actually told customers
What it cost
"DoNotPay made claims that its AI service could substitute for the professional expertise of a human lawyer… but [could not] substantiate those claims."
— U.S. Federal Trade Commission, on the DoNotPay settlement (January 2024)
Three review criteria that would have caught this
Each criterion below maps to a real review task you can configure in the sample builder. A certified reviewer checks every AI-generated marketing claim against these before it ships.
Substantiate all marketing claims with evidence
Every superlative, guarantee, or "proven" assertion must cite a study, benchmark, or dataset. If the AI writes "world's first" or "guaranteed to win," the reviewer blocks it until evidence is attached.
Verify "AI can do X" statements against actual capabilities
Any claim that the AI replaces, beats, or substitutes for a professional must be checked against measured capabilities. "Can replace an attorney" only ships if competence testing exists for the specific task.
No unauthorized practice of law claims
Copy must not state or imply the service practices law, gives legal advice, or stands in for a licensed attorney. These claims are blocked outright — they invite regulatory action regardless of capability.
Paste any AI marketing copy. See what gets flagged.
This is a simplified version of what our reviewers see. Paste product copy (yours or a competitor's) and run the check. The criteria above are applied automatically.
Don't wait for the FTC
Every unsubstantiated AI claim is a regulatory liability. Put certified reviewers between your marketing and the public. 50% off your first $10 — live in under 5 minutes.