110 incidents on record · 2026 Headlights Incident reports by Ellie Harris · Melbourne
10 new this week Library last updated 27 July 2026
← The incident library
HD-INC-098
Technology · United States · 2024 · Overstated capability

DoNotPay marketed a 'robot lawyer', but the FTC said it never checked whether the AI matched a real lawyer, and fined the company and barred the claim

By Ellie Harris · Filed Marketing claims from about 2021; FTC complaint 25 September 2024

Alleged: DoNotPay developed or deployed the AI system implicated in this incident. Details are drawn from public reports; parties are presumed innocent of any wrongdoing not established by an official finding.

DoNotPay marketed a 'robot lawyer', but the FTC said it never checked whether the AI matched a real lawyer, and fined the company and barred the claim

What happened

It was reported that in 2024 the United States Federal Trade Commission took action against DoNotPay, a company that had marketed itself as “the world’s first robot lawyer”, alleging that its product failed to live up to its claims as an adequate substitute for the expertise of a human lawyer. According to the FTC, DoNotPay told consumers the service would let them “sue for assault without a lawyer” and “generate perfectly valid legal documents in no time”, and that the company would “replace the 200-billion-dollar legal industry with artificial intelligence”. The catch, the FTC alleged, was the evidence: the company “did not conduct testing to determine whether its AI chatbot’s output was equal to the level of a human lawyer”, and “did not hire or retain any attorneys” to check the quality of its legal features. A separate feature that claimed to scan a small business’s website for hundreds of legal violations from its email address alone was, the FTC alleged, also not effective.

The action was reported to be one of the first in Operation AI Comply, the FTC’s sweep against overstated AI claims. Under the order, finalised in early 2025, DoNotPay agreed to pay 193,000 dollars, to stop advertising that its service performs like a real lawyer unless it has evidence to back it up, and to notify consumers who had subscribed between 2021 and 2023 about the settlement. DoNotPay settled without admitting or denying the allegations. Its founder’s company was reported to have told reporters it was “pleased to have worked constructively with the FTC to settle this case”, said it had not admitted liability, and described the complaint as relating to “a few hundred customers some years ago” using services that had “long been discontinued”.

What an auditable version would have shown

A claim like “as good as a human lawyer” is not a slogan, it is a testable proposition, and the FTC’s allegations focused on the absence of substantiated testing behind it. An auditable version would make the proof a matter of record: results from measuring the AI’s output against the standard it claims to meet, and a record of who, with the right qualifications, reviewed that output before the claim went out. Those records speak directly to the question the FTC had to ask, does the product do what it says, and put the burden of proof where it belongs, on the company making the claim, before a consumer relies on it rather than after a regulator investigates.

Where the gap was

A high claim was made about a professional-grade service, and, on the FTC’s allegations, no competent evidence measured whether it was true and no qualified professional had reviewed it. A VerificationGate holds a capability claim to independent evidence before it goes out, so “performs like a lawyer” has to be shown, not just asserted. A MetricRecord captures how the AI actually performs against the benchmark it invokes, as a standing number a company, a regulator or a customer can read, rather than an unmeasured marketing line. The failure, as the FTC framed it, was not that the tool had no value, but that no substantiated record connected the promise to proof.

What governance should have looked like

When a product claims to stand in for a trained professional, in law, in medicine, in anything where being wrong carries real cost, the claim should carry a higher burden of proof than an ordinary advertisement. Before it is made, the capability should be measured against the standard it invokes and reviewed by someone qualified to judge, and the results kept where they can be checked. What this case, and the wider sweep it was part of, comes down to is simple: sounding confident about an AI is not the same as having tested it, and a company should be able to back the claim up before it asks the public to rely on it.

Failure Pattern: an AI service was marketed as matching a professional’s expertise without substantiated testing against that standard or review by qualified professionals.

Governance Principle: a claim that an AI performs at the level of a trained professional must be backed by testing against that standard, and by qualified review, before it is sold or relied on.

The reference implementation of VerificationGate and MetricRecord is open source. It lives at github.com/saffronandindia/headlights-oss, Apache 2.0 licensed and free to install. The repository is public now.

Sources

The mailing list

Fresh incident reports every week. One email to match.

We add new incidents to the library regularly, and send a single short email each week with what's new. The library stays free and open; this is just how you keep up with it.

No tracking. Unsubscribe in one click.

The record

An auditable system would have produced a signed, tamper-evident record the moment this happened: what the system did, the version that did it, the basis it acted on, and the action taken, and DoNotPay could have produced it on demand.

This is the record the system as deployed did not produce in a signed, auditable form.

What this teaches
Capture what happened when it happens
What the system did, the version that did it, the basis it acted on, and the action taken, recorded at the moment, not reconstructed after.
Sign it, so no one has to trust the record-keeper
A tamper-evident entry. Edit it later and the signature breaks. The record does not ask for the benefit of the doubt.
Make it verifiable by anyone
A court, a regulator, a customer's lawyer can check the record themselves, without taking the company, or us, at our word.

Headlights summarises publicly reported AI incidents. All summaries are independently written, attributed to their original sources, and intended for research and educational purposes. Allegations are identified as such until established through official findings.

Last reviewed June 2026. This report is based on the sources listed above and reflects information available at the time of review; later developments may not be captured. Where a person is described as charged with or alleged to have done something, that allegation is unproven unless a conviction or a court or regulatory finding is stated. Headlights publishes journalism and commentary, not legal advice.

Want to write back?

Direct to my inbox.

ellie@useheadlights.com →