The AI Reckoning for Lawyers Is Here, and It’s Not a Quiz
The profession is not short of AI chatter; it is short of judgment, and the first lawyers to pay for that gap are already appearing in sanctions orders.
An AI-assisted editorial, reviewed by a human before publishing. It reasons over our own tracker data (and cited context) — a point of view, not legal advice.
The legal profession has a simple choice to make, though it keeps pretending otherwise: use AI with discipline, or let it use you with embarrassment. The sanctions are not theoretical. Courts have already punished lawyers for hallucinated citations, fabricated authorities, and AI-generated briefs that wandered off the map. That is not a future ethics problem. That is a present-tense malpractice smell.
And while some firms are building internal systems, training lawyers, and putting real governance around the tools, plenty of others are still behaving as if this were a novelty item to be tested after hours. The mismatch is the story. AI is moving into law whether a firm has a policy or not. The only question is whether the lawyer using it knows enough to verify the work before signing their name to it.
The machine does not bear your signature.
The risk is not that AI is smart. It’s that lawyers get lazy.
The sanctions tell the story plainly. Courts have issued show-cause orders, reprimands, monetary fines, and published cautionary decisions after lawyers filed papers with fabricated citations or hallucinations. One federal magistrate judge in North Carolina formally reprimanded a fired federal prosecutor for knowingly filing an AI-generated brief with fabrications and fake case citations. California appellate attorneys were sanctioned for briefs containing fabricated citations created by generative AI tools. If you want a professional-grade reminder that the machine does not bear your signature, there it is.
This is why the old excuse about “just using it for a first draft” is not a defense; it is a confession of poor supervision. A brief is not better because a machine wrote a few elegant sentences around nonexistent law. It is worse. The court does not care that your software was enthusiastic. The court cares that the citation was invented.
The opportunity is real, but it belongs to the organized lawyer.
The useful side of AI in law is not mystical. It is operational. Thomson Reuters’ CoCounsel Legal, Harvey, Lex Machina, and RelativityOne all sit at the top of the tools catalog, and the uses are familiar enough: research, analysis, drafting, document review, litigation intelligence, e-discovery. Firms and legal departments are also building beyond off-the-shelf products. Descrybe released an open connector for custom workflows. Watts Law launched an AI case-qualification platform that still keeps every case subject to lawyer review. Kirkland & Ellis has earmarked $500 million of its revenues to develop its own AI platform. Cleary Gottlieb bought an AI firm and brought engineers in-house. None of that sounds like a fad.
The message for a working attorney is not “replace yourself.” It is “stop pretending efficiency is optional.” Corporate legal departments are being told to reinvent talent, data, and workflow strategies. Gartner says legal teams already treat AI as a top priority and predicts companies will double legal technology budgets by 2028. In-house teams are acting accordingly: Microsoft’s legal and compliance organization is using Harvey, and other major legal teams are building AI policies, governance frameworks, and internal programs. If the client side is changing, the billable hour is not going to save anyone from the consequences.
The real divide is not between believers and skeptics. It is between supervised and sloppy.
There is a temptation to frame all this as an adoption race, but that is too neat. The profession’s actual reckoning is more basic: who can prove they controlled the tool, and who cannot. AI-related sanctions are already being tracked, with California leading the states named in the data. That should be enough to cure anyone of the fantasy that “everyone is doing it” is a usable compliance policy.
A lawyer who uses AI well will move faster, search better, and spend less time doing labor that clients do not want to fund. A lawyer who uses it badly will create a record that judges, opponents, and bar counsel can read aloud. The profession does not need more faith in AI. It needs adults in the room.