A law firm had an associate leave. Instead of hiring a replacement, they gave the work to AI. Costs dropped 27%. Profits went up. They published a case study about it.
Think about that for a second. A person spent three years in law school. Passed the bar. Spent years building expertise. And the firm’s honest assessment, after running the numbers, was that an AI subscription was a better use of the money.
This is not a hypothetical. This is Above the Law, February 2026.
What’s Actually Happening in Law Right Now
None of the AmLaw 100 firms say they’re planning to reduce headcount. That’s the reassuring headline. But none of them are replacing departed associates either. The headcount stays flat not because AI isn’t doing the work, but because the work is getting done with fewer people through attrition.
You don’t fire lawyers to replace them with AI. You just stop hiring new ones. The outcome is the same. The optics are cleaner.
Contract review, due diligence, legal research, first drafts of briefs. All of it is now AI-assisted at minimum and AI-replaced at maximum in most major firms. The junior associate who used to bill 80 hours reviewing documents no longer exists. Neither does their salary. Neither does their career path.
The Accountant Problem Nobody Expected
Half of accountants are reporting client losses because those clients took AI tax advice and it was wrong. Half. That’s not a rounding error. That’s a profession in the middle of a crisis it didn’t create and can’t fully control.
92% of accountants say public AI tools should be regulated when giving financial advice. The irony is that the people arguing hardest for AI regulation in accounting are the same people watching AI take their clients. They’re not anti-technology. They’re watching people get hurt by bad AI advice and spending thousands of hours cleaning up the mess.
The AI isn’t replacing accountants by being better. It’s replacing them by being cheaper and convincing enough that clients try it first, get it wrong, and sometimes never come back.
Congress Noticed the Doctor Problem
In March 2026, lawmakers introduced a bill specifically to stop AI chatbots from impersonating doctors, lawyers, and licensed professionals. The bill exists because people are already treating AI responses as professional advice. Because the chatbots sound authoritative. Because they’re available at 2am when the anxiety is worst and the doctor’s office is closed.
The fact that a bill had to be introduced means the problem is already real enough to require legislation. You don’t write laws about hypothetical behaviors.
The Thing Nobody Wants to Say
Professional licensing exists for a reason. Doctors, lawyers, and accountants go through years of training and pass rigorous exams because the stakes of getting it wrong are high. A bad legal brief costs someone their case. Bad tax advice costs someone their savings. Bad medical advice costs someone their life.
AI gets these things wrong in ways that are hard to detect because the wrong answer sounds exactly like the right one. Confident. Well-structured. Authoritative. The hallucination problem isn’t a bug that will be patched. It’s a property of how these systems work.
But the economics don’t care about the error rate. They care about the average outcome versus the cost. And right now, for a huge number of routine professional tasks, the average AI outcome is good enough, and the cost difference is enormous.
The professions aren’t going away. But they’re going to look very different by the time the dust settles. And the people who are three years into a law degree right now deserve to know that before they finish.
[…] often struggle to get timely appointments for what are essentially administrative renewals. The cost of professional time is already being used to justify AI displacement across every field, and in healthcare that argument has genuine force because access disparities kill […]