The Last Class of Paralegals

Priya Menon finished her paralegal certificate at Maricopa Community College in December 2024. She was thirty-one, the first in her family to work in a professional office, and she had spent $8,400 on the program. By March, she had sent 140 applications. She got four callbacks — all of them rescinded before she started, as the firms froze hiring to "evaluate new efficiencies." She is now working at a Target in Tempe, on the early morning stock shift, which starts at 4 a.m.

I heard about Priya from a workforce counselor in Phoenix who asked me not to use her last name. The counselor had seen four other paralegal graduates in the same situation in a single quarter, and she was trying to figure out what to tell the next cohort. She didn't have an answer. I didn't either, and I kept waiting for one to arrive while I was reporting this piece. It didn't.

---

The legal industry's adoption of AI document review and research tools has been faster and quieter than almost any comparable shift in white-collar work. Tools like Harvey, CoCounsel, and Lexis+ AI can now cut legal research time by more than half, according to testing by several large firms — work that used to flow to junior associates and paralegals almost automatically. The tools don't replace partners. They don't replace experienced litigators. They replace the people who were learning.

That's the part that rarely gets mentioned in the productivity announcements. When a law firm says it processed twice as many contracts this quarter with the same headcount, the math is clean and the press release is flattering. What it obscures is that "same headcount" means the senior people who were already there — the ones who didn't need to be hired, who already had their credentials confirmed, their bar admissions done, their names on the letterhead — while the entry-level positions that normally would have been added weren't, the pipeline dried up, and nobody scheduled a press conference about it.

The Credential That No Longer Opens the Door

For decades, a paralegal certification or a junior associate position at a mid-size firm was a reliable way in: into stable income, into professional networks, into the possibility of law school later if you wanted it. It was one of the few paths that didn't require a four-year degree at a name-brand institution. A class-mobility mechanism, working quietly and mostly well.

The people on that path now are in an awkward place. The credential is real. The skills are real. The demand for those skills, performed by a human at entry wages, has contracted sharply. The Georgia Bar Association's 2025 workforce survey found that paralegal job postings in the state dropped 34 percent year-over-year. Arizona's numbers are similar. These aren't laid-off veterans with severance packages and savings. They're people who just paid for a certificate and can't get the first job.

What makes this shift hard to see is that it doesn't produce dramatic headlines. No mass layoffs. No plant closings. No press conferences. The jobs simply don't open. The cohort that graduated last winter goes mostly unnoticed, because they never formally entered the workforce they were aiming for.

What Gets Learned in the Doing

There is something else worth saying, beyond the economic argument: the work these junior employees did wasn't only useful for output. It was useful for formation.

A paralegal who spends two years pulling documents for discovery learns how cases are built. A junior associate who drafts the same motion forty times starts to understand how argument works — not as a formal exercise but as a felt sense of what holds and what collapses under pressure. The inefficiency of that learning process was, partly, the point. You made mistakes in low-stakes contexts. A partner caught them. You got better.

If that process is compressed or skipped — if AI produces the first draft, the summary, the research memo — the learning still needs to happen somewhere. No one has figured out where.

I caught myself wanting to write that the partners who use these tools were themselves, once, the people learning the slow way, and that they sometimes forget that. True enough. But the more uncomfortable thing I noticed is that the forgetting is convenient. The partners don't lose anything by skipping the formation stage for the next generation. The firms are more efficient. The press releases are good. The cost is paid entirely by people who haven't started yet.

The Counselor's Question

The workforce counselor in Phoenix told me she had started advising her clients not to pursue the paralegal track unless they had a backup plan. She said this with visible discomfort, because she knew what the alternative paths looked like: customer service, logistics, healthcare support roles that were harder and paid less.

What she was really saying is that she didn't know how to promise the future she had been promising. The credential used to be a reasonable bet. Now it was something closer to a gamble.

I keep thinking about Priya at Target at 4 a.m., and about what it means to be the last class of paralegals — not in the sense that the profession disappears entirely, but in the sense that it stops being a door and becomes something rarer, something you need connections or luck to reach. That's not an accident and it's not temporary. The firms made a calculation, the tools delivered, and the people holding freshly printed certificates are the ones who absorbed the loss. A smaller change than a plant closing, yes. Also a more permanent one. And one with nobody's name on it.