Share
For nearly fifty years, the economics of a BigLaw associate class have rested on a simple, well-understood assumption: junior lawyers spend their first several years doing the unglamorous work—first-draft research memos, document review, redlining routine contracts—and in doing so, they generate billable hours and teach themselves how to practice law. Those two functions, billing and training, have always run on the same track. That track has been unravelling for years now, and early data out of 2026 suggests that it's happening faster than most students or junior associates understand.
The Numbers Behind the Shift
Beginning with hiring math, a SurePoint study covered by The Global Legal Post found that lateral associates accounted for 51% of Am Law 200 associate hires last year—the third time in five years that laterals have outnumbered entry-level hires. Associate lateral hiring rose 12% year over year in the same period, while entry-level hiring “flatlined.”
Additionally, BCG Attorney Search's 2026 Legal Talent Movement & Law Firm Headcount Report says that 86% of large firms plan to grow their overall associate ranks through 2027—but only 35% plan to grow first-year class sizes specifically. Firms want more lawyers. They just don't want more first-year associates.
Why This Is a Training Problem
"AI is doing junior work now" is inadequate to describe what’s happening. The tasks AI has already begun to absorb—document review, first-pass drafting, legal research—were never just busywork. They were also a huge part of how junior associates were trained, and allowed to practice and develop their craft for themselves.
Axios spoke with a few professors about this tension. Nik Guggenberger from the University of Houston Law Center put it bluntly: junior work "has always served two purposes: billing and training," and "if more and more of that work that trains junior associates is being automated, then there's no real material anymore for them to train on." David Freeman Engstrom, professor of law at Stanford, described firms as racing to "extract the knowledge of their lawyers" and embed it directly into AI workflows and client-facing tools. This may be efficient in the short-term, but ignores the problem of training up the next generation of lawyers.
The same piece states that the "leverage model" of law firm economics—a small number of partners sitting atop a much larger base of billing associates—is facing a “structural threat” if the work that justified that pyramid shape simply isn't there anymore. There is a real risk of “creating lawyers who can supervise AI outputs” without being able to judge whether those outputs are correct.
How Firms Are Responding
Some law firms are already treating AI as core infrastructure. A&O Shearman announced it would roll out agentic AI tools built with Harvey aimed at complex legal workflows, for use both internally and as a product sold to clients and other firms.
BCG's 2026 report found that through the first nine months of 2025, equity partner headcount at large firms actually declined 0.5% even as total lawyer headcount grew — meaning all headcount growth came from salaried, non-equity lawyers. Leverage (the ratio of associates to partners) rose 4.3% over the same stretch. Firms are getting bigger at the base and thinner at the top of the traditional partnership track, which is close to the opposite of what "more headcount" used to signal.
What This Means For Current Students
For students evaluating offers, this changes what a firm's prestige rankings can tell you on their own. Two firms with similar prestige rankings—or even Quality of Life rankings in technology and innovation, or associate training—may be making very different choices on how they structure junior work, and those choices will shape your first few years far more than the name on the letterhead. It's worth asking, directly, how a firm is restructuring associate training around AI tools—not just what the tools are, but who still gets to touch the underlying work and learn from either doing or properly vetting it.
The same Axios reporting offers an alternative view courtesy of Tiffany Tucker, assistant dean for career development at the University of Houston Law Center: students with AI skills are becoming "the more attractive candidates," while those without them risk being "left behind." Engstrom's framing is worth considering too—the lawyer of the future isn't a document reviewer, but closer to a "symphony conductor" who pieces together AI output, data, and legal scenarios into something clients can rely on. That's a different skill than the one the old apprenticeship model was built to teach, and it's one you'll likely have to build more deliberately rather than simply absorb by osmosis while billing hours.
A Caveat
None of this means the answer is obvious, even to the firms making these choices. Nobody—including the firms furthest along on AI adoption—has adequately answered the question of where skill comes from if the traditional training ground disappears.
The Bottom Line
Pay scales are up. Total headcount is up. By almost every traditional metric, BigLaw looks healthy. But underneath those numbers, the shape of what a legal career actually looks like in changing—and 2026 appears to be the first year several independent data sources agree the shift is visible, not just anticipated. If needed, we may revisit this topic as more firms report full-year 2026 hiring figures later this year.
Share
