Workforce Transformation
How AI is changing the legal workforce: hiring, training, and the people.
Opening Argument
One argument, drawn from the week to date.
Four items this week point at the same shape, not just a smaller one. Above the Law, citing a new Citi survey, reported first-year associate hiring down 7.5 percent for the class of 2025, and Citi's Gretta Rusanow said roughly two-thirds of large firms expect a smaller first- and second-year population still by 2028. That is the base of the pyramid narrowing.
At the same time, Latham & Watkins is recruiting for Legal Engineering, Knowledge Transformation, and AI Services attorney roles paying $180,000 to $400,000, a track with no seat on the traditional ladder. FairPlay Law launched this week as a fixed-fee, AI-native employment firm built to never need that ladder at all. Two new tracks are opening beside the one that is narrowing, and the class still coming in, 96 percent of whom used AI this summer against 80 percent last year, cannot yet know which shape it is training for.
The Implication
Track headcount by role type, not the associate count on the letterhead. A stable topline number can hide a workforce splitting into a narrower traditional track and a smaller, differently paid technical one.
The Findings
What moved in practice this week, and what each one tells you.
First-year associate hiring at large firms fell 7.5 percent for the class of 2025, and two in three firms expect it to shrink further by 2028.
Citing a new Citi law firm advisory survey, Above the Law reported on 11 September that first-year hiring fell 7.5 percent from the class of 2024 to 2025. Citi's Gretta Rusanow said the pattern points toward smaller first- and second-year populations at a majority of large firms by 2028.
Why it matters The 2028 figure is a survey of firm expectations, not a locked outcome. The 7.5 percent drop is the measurable part. Check it against your own firm's actual incoming class rather than assuming an even effect across practice groups.
Latham & Watkins is recruiting for AI-specific attorney roles paying up to $400,000, titles that did not exist a decade ago.
Artificial Lawyer's 14 September review of Latham's careers page found the firm hiring for Legal Engineering Attorney, Knowledge Transformation Attorney, and AI Services Attorney positions, at $180,000 to $400,000. The piece notes that in 2016, many firms had nobody on staff with the word “innovation” in their title.
Why it matters These roles sit outside the traditional associate track, with their own pay bands and no obvious partnership path yet. Ask whether your AI hires are a durable career line or a project function likely absorbed back into existing roles once the tooling settles.
96 percent of this year's summer associates used AI on the job, up from 80 percent last year.
Law360 Pulse's second annual summer associate survey, 486 responses collected 1 July to 18 August, found AI use ranging from 83 percent at smaller firms to 99 percent at the largest. Eighty-seven percent of firms encouraged AI use and 9 percent required it; maximum stipends at the largest firms topped $57,000, up from $50,000.
Why it matters This is a self-selected respondent pool, not a census of every summer class. Set it beside the hiring data above: firms are training this cohort harder on AI while narrowing the seats waiting for them afterward.
A new AI-native employment law firm launched this week, built to never need the associate ladder at all.
FairPlay Law, founded by David Perla and Sanjay Kamlani and reported 16 September, serves individual employees on job offers, equity, severance, and workplace disputes, offering a free AI-generated document review ahead of a paid, fixed-fee consultation. It pairs with FairPlay Global, a separate management services organization, built to work within state rules on law firm ownership.
Why it matters One new firm is not a track record. What it signals holds regardless: a fixed-fee, technology-first model can launch around a leverage pyramid it never has to build, competing for the same junior talent traditional firms are training.
A pricing survey found 27 percent of firms cut billable hours on certain tasks after deploying AI, and a third of partners are not comfortable telling clients.
BigHand's pricing survey, reported by Artificial Lawyer on 16 September, found 31 percent of firms that analyzed their own data saw improved profitability per matter, 30 percent saw increased throughput, and 27 percent saw reduced billable hours on specific tasks. Separately, 35 percent of partners said they were not comfortable discussing their firm's AI use with clients.
Why it matters The reporting we read did not disclose BigHand's sample size or selection method, so treat this as a signal from firms already measuring, not a market average. The discomfort figure is the workforce point: fewer billable hours on a task changes what doing it is worth to bill for, ahead of any compensation model catching up.
The Practice
This week's deep read on AI in practice.
The leverage pyramid has described law firm economics for decades: a broad base of first- and second-year associates doing volume work that senior lawyers price and review. This week's numbers do not show that base simply thinning in the same shape. They show two different bets being placed in the same market at once, and only one looks like the pyramid getting smaller.
Above the Law's report on the Citi survey put first-year hiring for the class of 2025 down 7.5 percent, with roughly two-thirds of large firms expecting a smaller first- and second-year population by 2028. That is a bet that entry-level work volume keeps falling. Artificial Lawyer's review of Latham's recruiting page is a different bet: Legal Engineering, Knowledge Transformation, and AI Services attorneys at $180,000 to $400,000, titles that answer to no existing seniority ladder. One bet shrinks the base. The other opens a column beside it.
96%
Share of this year's summer associates who used AI on the job during their program, against last year's cohort.SOURCE · LAW360 PULSE, 2026 SUMMER ASSOCIATES SURVEY
On the Methodology
486 responses collected 1 July to 18 August 2026, Law360 Pulse's second year running this survey of its own audience of summer associates. It measures self-reported use among students who responded, not a random sample of every program in the country.
These are the students who become the first-year class the Citi survey says firms expect to keep shrinking, training more intensively than any class before them on tools that, by their own firms' hiring signals, may be doing some of the work their jobs used to justify.
FairPlay Law's launch is a small data point alone, one new firm with no track record yet. But it tests a structural question live: an AI-native, fixed-fee firm built around David Perla and Sanjay Kamlani's prior ventures does not need a leverage pyramid to be profitable, since it was never designed to sell associate hours. At scale, that model competes for the same junior talent the traditional pyramid is narrowing, without building the ladder those associates expected to climb.
None of this week's items confirm AI is actually replacing the work junior associates used to do. BigHand's 27 percent reduced-hours figure is a correlation, not proof the two move for the same reason, and the survey did not disclose its sample or selection method. What the items do establish is a set of moves that only make sense if firms believe the substitution is coming: a narrower entry class, a new high-paid technical track, and an outside competitor betting it will not need juniors at all. Watch whether each decision rests on a measured read of what AI is doing today, or an assumption borrowed from someone else's headline.
Exhibit AI
When AI goes wrong in the legal profession. Not to assign blame, but to learn the lesson before it is ours.
In February, Baker McKenzie confirmed plans to cut up to 1,000 business services roles globally, roughly ten percent of that workforce, across research, marketing, and secretarial functions in London and Belfast. A firm spokesperson said the review was aimed at “rethinking the ways in which we work, including through our use of AI.” One affected employee called the decision “short sighted.”
The ABA Journal later asked AI and legal operations consultants whether the justification held up. Cat Casey of Masters AI called AI “a very convenient boogeyman to blame,” arguing that changing business needs, not verified AI capability, were the likelier driver, “a very good excuse for business necessary consolidation.” Jennifer Johnson of Calibrate Strategies described AI as an accelerant that exposed inefficiencies already there, not a technology performing the cut roles' work.
Notice the gap the critics point at. The firm named AI as a reason before anyone could show it was doing the eliminated work today, not eventually. That is the same gap this week's hiring numbers are quietly testing, where expecting AI to substitute for junior work gets treated as interchangeable with proof that it already does.
The Lesson
Cite AI as a staffing rationale only for work you can show it is doing today, not work you expect it to do eventually. That gap is where every AI-washing critique lands, and it sits behind this week's shrinking entry classes until someone measures it.