DoNotPay marketed itself as "the world's first robot lawyer" and promised to fight parking tickets, cancel subscriptions, and eventually handle real legal matters. Then the FTC sued, state bars took notice, and the company quietly retreated from its legal claims. Here's what actually happened — and what it means.
DoNotPay launched in 2015 as a chatbot that helped users fight parking tickets. The concept was simple and genuinely useful: translate the bureaucratic process of contesting a ticket into a guided conversation and produce the correct form letter. It worked well enough in that narrow context.
Over the following years, the company expanded aggressively into new areas — disputing bank fees, canceling subscriptions, claiming refunds, writing demand letters for landlord-tenant disputes, and eventually marketing itself as a legal AI that could handle "any legal issue for $36/year." The founder made increasingly bold claims about replacing lawyers, appeared on major media, and raised substantial venture capital.
The marketing explicitly positioned the product as a lawyer substitute — not as a document helper or information resource, but as something that could handle your legal situation the way a lawyer would.
State bar associations regulate who can practice law. Providing legal advice — not just information, but specific guidance on what to do in a legal situation — is generally considered the practice of law. When DoNotPay moved beyond document templates into legal strategy and advice, it entered territory that state bars consider reserved for licensed attorneys. Several state bars raised concerns.
In 2024, the FTC sued DoNotPay, alleging that the company had made deceptive claims about its AI capabilities — specifically that it had not adequately tested the product to verify its claims about legal expertise and outcomes. The company settled for $193,000 and agreed to stop making certain claims about its products' capabilities without substantiation.
In early 2023, DoNotPay announced it would pay $1 million if a defendant used its AI via earpiece to argue a Supreme Court case. The plan drew immediate criticism from bar associations, legal ethics experts, and the courts themselves — several state bars issued statements that using a concealed AI device to receive real-time legal coaching could constitute unauthorized practice of law and potentially fraud on the court. The stunt never happened.
This episode exposed a fundamental tension at the center of the AI legal tool space: the difference between helping someone prepare for court and impersonating counsel in court.
Setting aside the overpromising, DoNotPay identified a real problem: most people facing routine bureaucratic and legal situations have no access to professional help and no understanding of their options. The parking ticket tool, the subscription cancellation tool, the fee dispute templates — these were genuinely useful. Simple, automatable tasks that didn't require legal judgment.
The failure was in the leap from "document helper" to "robot lawyer." That leap required the product to actually deliver on legal competence it didn't have — jurisdiction-specific knowledge, accurate case analysis, reliable legal reasoning. When users relied on DoNotPay for matters beyond its actual capabilities, the results were predictably bad.
The DoNotPay story is a case study in what happens when a legal AI product overpromises on capabilities, under-delivers on actual legal knowledge, and doesn't take seriously the distinction between information and advice.
We are not trying to be a robot lawyer. We are not claiming we can handle your legal situation. We are explicit that we are not a law firm, that nothing we produce is legal advice, and that no attorney-client relationship is created by using our platform.
What we are: a legal intelligence platform that processes your specific case documents and returns analysis grounded in those documents — charge elements, evidence weaknesses, motion opportunities, judge intelligence. The output is information you can act on, not advice that creates liability for us or false confidence for you.
See: why a generic AI isn't built for your case, and what we are and what we're not.
Be My Own Attorney is built around what AI can actually do well for defendants: process documents, surface evidence weaknesses, and provide judge intelligence. No overclaiming. No false confidence. No courtroom earpiece stunts.
Join the Waitlist