Will AI Replace Lawyers? An Honest Take

Goldman Sachs published a report in 2023 saying AI could automate 44% of legal tasks. Headlines screamed about the end of the legal profession. A client forwarded it to an attorney we know and asked if he'd still be in business in five years.

He's been practicing plaintiff PI for 22 years. He laughed.

Three years on, here's what we actually know, and where the Goldman Sachs report gets it right.

What AI Actually Changes in Plaintiff Work

Let's be honest about what AI does well in a plaintiff practice today, and what it actually unlocks for firm growth:

Medical record review. AI processes hundreds of pages of medical records in minutes, producing chronologies, flagging bad facts, identifying missing records, and calculating damages automatically. What used to take a paralegal 30–40 hours now takes less than an hour. Attorneys who aren't buried in records can be more selective about which cases they take, and more thorough on the ones they do.

Intake and case evaluation. AI screens inbound calls, extracts key facts, scores leads, and surfaces the strongest claims automatically. Frontier Law Center cut their intake and case evaluation process from three to four hours down to thirty to forty minutes, and their intake conversion rate went from 10% to 35%. That's a fundamentally different revenue equation.

First-draft generation. Demand letters, discovery responses, motions, deposition outlines: AI produces first drafts that attorneys edit and refine, calibrated to the firm's own tone and formatting. Hours saved per document, multiplied across hundreds of cases, add up to attorneys spending more time on the work that actually moves cases forward.

Document review and case intelligence. AI handles the volume that would bury a small firm, sorting discovery, identifying relevant documents, flagging privileged material. The most sophisticated implementations go further: running nightly scans across a firm's entire caseload to surface hidden value, undiagnosed conditions, and coverage gaps that human review misses.

Legal research. Finding relevant cases, summarizing holdings, answering case-specific questions from uploaded documents. AI is faster than manual research for most tasks, freeing attorney time for analysis and strategy rather than retrieval.

The numbers bear this out. Michael D'Isola, a 40-year litigation veteran at Jeffrey Glassman Injury Lawyers in Boston, recently worked a wrongful death case that settled for $9 million. He used AI to summarize deposition transcripts. "What would take two hours per deposition, Eve did in 20 minutes each," he said. "That case had 10 to 12 depositions. What would usually take a week, I did in one afternoon."

Add it up and yes, 44% of legal tasks is a reasonable estimate for what AI can handle. Goldman Sachs wasn't wrong about that number.

They were wrong about what it means.

What AI Can't Touch

Here's what's missing from every "AI will replace lawyers" article:

Client relationships. A woman walks into your office. She was rear-ended by a commercial truck six months ago. She's in chronic pain. Her employer is pressuring her about missed work. She's scared. She starts crying in your conference room.

No AI handles that moment. No AI looks her in the eye and says "I'm going to fight for you, and here's exactly how." No AI makes her feel heard, protected, and represented. That moment, when a human being trusts another human being with their pain, is the foundation of every plaintiff case.

Courtroom strategy. Try putting an AI in front of a jury. Watch it read body language during voir dire. Watch it adjust closing arguments on the fly because juror #4 has been checking out since lunch. Watch it make the split-second decision to go hard on cross-examination or back off because the witness is about to cry and the jury is sympathetic.

Trial work is performance. It's psychology. It's instinct built on decades of experience. AI isn't close.

Judgment calls. Should you take this case? Is this client telling the truth? Is this insurer going to lowball and force litigation, or can you settle pre-suit? Should you demand $500K or $1.2M? Should you accept the mediation offer or roll the dice at trial?

Every one of these decisions requires experience, intuition, and contextual understanding that AI simply doesn't have. AI can provide data to inform these decisions. It cannot make them.

Negotiation. The adjuster knows you'll take this to trial. That changes the number. Your reputation precedes you. The defense attorney knows your trial record. None of that exists in an AI model. Negotiation is fundamentally human, fundamentally relational, and it's where most plaintiff cases are won or lost.

Ethics and judgment under pressure. When you discover something in the records that hurts your client's case, what do you do? When opposing counsel makes an improper offer, how do you respond? When your client wants to lie? These aren't task-completion problems. They're moral and professional judgment calls.

Will AI Replace Lawyers? Here's What's Actually Happening

Here's what's actually happening in the firms that are using AI aggressively: they're not firing attorneys. They're handling more cases with the same team.

Mike Morse Law Firm, with 250 legal professionals and over 60 attorneys, is one of the largest personal injury firms in Michigan and one of the most aggressive AI adopters in the country. When Morse first rolled out AI tools, his team was skeptical. The number one barrier, he says, was fear: "Employees have been doing it one way for a very long time, and they're fearful of losing their jobs."

Today, the dynamic has completely reversed. Attorney Natalie Jaboro uses AI roughly 75 times a day: drafting demands, reviewing medical records, and supporting live calls with adjusters. Her caseload capacity has doubled, minimum. When Morse asked his team how they felt about the tools, the response surprised even him: "They've told me that if I get rid of Eve, they're quitting."

The attorneys who were most afraid of being replaced by AI are now the ones who say they can't practice without it.

This is the 44% story that Goldman Sachs missed: automating 44% of tasks doesn't eliminate 44% of lawyers. It makes the remaining percentage of their work, the high-value work, more scalable.

The attorneys are spending more time in court, more time with clients, more time negotiating. The work that AI can't do. The work that actually determines outcomes.

Jennifer Horn runs operations at James Scott Farrin in Durham. She described a recent example: one of the firm's department heads blocked off two full days to prepare for a defendant deposition. He tried Eve instead and finished in a few hours. "Think about how much more time he was able to give to other cases," Horn says. "Before, you had a lot of administrative tasks. Now we can shift that to Eve and give everything we have to bring full value to the case and give a higher level of client care."

James Farrin, the firm's president, frames it as a philosophy: "Technology and AI is meant not to be between us and our clients, but behind us and our clients. It allows us to do a better job face-to-face, but I don't want it to be a substitute for the human client experience."

"This is not replacing lawyers or paralegals," Morse says. "These are tools to help us get better and help more people."

Who Should Actually Be Worried

If AI isn't replacing plaintiff attorneys, who is at risk?

Document review attorneys in the BigLaw model are in trouble. Contract attorneys billing $50/hour to review discovery documents are directly competing with AI that can do it faster and cheaper.

Forms-based legal services, including basic wills, simple incorporations, and uncontested divorces, are already being disrupted by AI tools. There's no strategic judgment required, so there's no moat.

Attorneys who refuse to use AI are the ones who should worry most. The firm across town that does use AI can handle more cases, respond faster, and operate more efficiently. You don't lose to AI. You lose to the attorney who uses it.

Mike Morse puts it plainly: "Law firms who don't take AI on right now are making a mistake. Their competitors are gonna pass them by."

The Five-Year View

In five years, the plaintiff bar will look materially different but not quite in the way the Goldman Sachs headlines suggested.

AI will handle most first-draft work, most medical record processing, and most routine research. That's table stakes. What changes more significantly is what attorneys do with the time they get back. The firms that use AI well won't just be faster; they'll be more selective about which cases they take, more thorough on the ones they pursue, and better positioned to build the client relationships that drive referrals and reputation. The ceiling on what a well-run plaintiff firm can accomplish, with the same headcount, goes up substantially.

The quality gap between firms that use AI well and firms that don't will widen every year. Client expectations will rise alongside it: faster turnaround, better communication, more transparency about where a case stands. Firms that can deliver that will grow. Firms that can't will find it harder to compete for the cases worth having.

None of that looks like "AI replaces lawyers." All of it looks like "AI changes what it means to run a plaintiff practice," and raises the floor on what clients should expect from theirs.