AI Won't Replace Me
Why I'm Glad I'm a Courtroom Attorney and Reflections on the Coming Upheaval
As an attorney, I’m part of a guild. The State Bar of Georgia controls entry into my profession. I carry a bar card, probably my most valuable possession. That little piece of plastic gives me the right to appear in court, so long as I remain competent, do my continuing legal education, and avoid serious misconduct. It also protects me from AI competition, because computers can’t conduct court appearances. Even if computers were physically capable of doing this, they’d still be prohibited—just like a plumber or a nuclear scientist would be. Only licensed attorneys can conduct court hearings for other people.
My practice revolves around these hearings. When someone gets a DUI or wants a divorce, they have to face a judge. I know how to talk to clerks and judges, and negotiate with other lawyers. I give people access to a fair hearing, where they can testify and/or present evidence, if they choose.
Most people prefer to avoid trial and resolve their cases by agreement. These negotiations are more complicated, and more emotional, than poker, because the goals are murky and often shifting. It’ll be a long time before any game-theory-optimizing algorithm handles a divorce better than a good human lawyer.
That’s why, for lawyers like me, AI isn’t the enemy. It’s a force multiplier. Any tool that helps me close cases faster or cuts down on clerical work makes me more efficient. With AI, I can keep money I’d otherwise spend on staff, in exchange for moderate licensing fees. The market for these tools is competitive, which means they’ll likely stay far cheaper than the labor they replace.
A huge swath of legal work is about to become obsolete, and that’s no tragedy. Lawyers are expensive, and that drives up the cost of everything from housing to healthcare to collecting child support. America would be wealthier if it had fewer lawyers and more affordable legal services. AI can help make that happen.
My first year out of law school, I clerked for the Honorable Lydia Taylor and Charles Poston. The core of my job was writing memos and drafting opinions. All of my work was reviewed. I was there to research and suggest—not to decide. It was better for the judges’ calendars (and my ego) when they followed my recommendations, but no draft of any length went through untouched. AI could easily do that kind of work. If I were a judge today, I’d rather have two clerks with AI than three without it. And AI is getting smarter—at least as quickly as my 11-year-old.
It can already handle most of what I did early in my career before I became a courtroom attorney. It can summarize documents, parse cases, and spit out solid first drafts. It is a patient and thoughtful editor that leaves the writer fully in charge. AI doesn’t get mad when I reject its suggestions. I don’t lose face when I accept its. There’s no tug of war. When using AI, you might have to swap in fake names to preserve confidentiality, but Ctrl+F works just fine for that and is quicker than arguing with a human editor or soothing one anothers’ feelings.
AI is already transforming legal research. It’s not a substitute for knowing the law, but a responsible lawyer can treat its output like a law clerk’s: something useful, not final. That’s still a massive time savings. When I worked as a staff attorney for the U.S. Court of Appeals in Atlanta, I was expected to screen four to six cases per month—and got extra time off if I screened more. An experienced attorney using well-trained AI could probably handle several times that caseload.
Job-specific legal AI still needs work, but when it comes, the need for paper-pushers will fall off a cliff.
It’s probably a bad time to go to law school. The demand for junior attorneys is about to crater. You’ll be swimming upstream. Still, a strong fish can swim upstream—and if you have a real vocation for the law, and you’re genuinely talented, there will still be a place for you.
Think about the cage matches that have existed in academia for assistant professorships. Landing a BigLaw associateship is about to look like that. Why? The biggest difference between a country lawyer’s practice and BigLaw is the ratio of courtroom time to back-office work. BigLaw work is much less court-facing. More of it occurs behind the scenes and can be seamlessly passed to AI without the judge knowing or caring. Many of those on the BigLaw bubble will pivot to small firm, courtroom work. They’ll be young, fast, and eager to build practices around AI from day one. More people will want to be courtroom attorneys.
Fortunately, law has always been gentler on age than most professions. I suspect that will continue. Judges tend to respect lawyers their own age more than young’uns. Established practitioners will be fine, but those trying to build something from scratch are in for a fight.
In 20 years, the best legal writers may be the ones who write the best AI prompts. Maybe there will still be a market for purely human prose—but it’s unlikely that market will include much legal writing. Legal writing is functional, forgettable, and rarely loved. Maybe courts will still require a human with a bar card to sign off before anything gets filed. Even so, the armies of clerical workers (including desk-bound attorneys) who’ve supported the profession are in for upheaval. Those with strong people skills will hang on. Those who master the machines will thrive.
Every redundant job will bring us closer to what the law is supposed to be about: the just and efficient resolution of disputes. If AI can lower the cost of lawyers, it will open the courthouse doors. It will empower clients to ask better questions, resist manipulation, and pursue realistic goals. I’m ready for that brave new world—a world where the scales of justice tilt a bit less toward the mighty dollar.