This fall, if you walk through the recruiting floor of any major Am Law 100 firm, you’ll notice that it’s quieter than it was. There are fewer offer letter stacks. Once teeming with first-years reviewing documents in borrowed conference rooms, the summer associate programs have become much smaller. Now that the numbers supporting that sentiment are known, they are difficult to ignore.
First-year associate hiring at Am Law 100 firms fell by almost 17% in a single recent fall cycle, according to the Thomson Reuters Institute Law Firm Financial Index. The drop was even more severe at Am Law 200 firms, averaging about 25%. According to NALP, the largest companies are offering their smallest summer classes since 2012, and median 2L summer program offers have reached their lowest level ever. These are not rounding mistakes. This is a structural contraction that is occurring more quickly than the industry is willing to acknowledge.
More than anything else, AI’s true operationalization within law firms is driving the change. Approximately 70% of lawyers report using AI on a weekly basis as of early 2026. This means that they are incorporating it into their daily workflow rather than just experimenting with it or attending webinars about it. Furthermore, first-pass document review, legal research, and first-draft contracts are the kinds of tasks that AI excels at and that have historically justified hiring large classes of junior associates. There is still work to be done. Simply put, a 25-person first-year class is no longer necessary to accomplish this. It’s getting more difficult to ignore the agency recruiting issue that legal scholars have been quietly bringing up. Businesses will need to develop a new apprenticeship system if AI eliminates the low-level repetitions that foster junior judgment. Otherwise, they run the risk of producing attorneys who can oversee AI outputs without having developed the intuition to recognize when those outputs are incorrect. That isn’t speculative. Businesses are currently creating this development gap in their hiring decisions this year. Axios
The largest AI-attributed workforce reduction in the legal sector to date occurred in February 2026 when Baker McKenzie eliminated between 600 and 1,000 business-services positions, specifically citing AI integration. Lawyers weren’t the main roles that were eliminated. These included document processing, coordinating research, and handling administrative tasks related to legal work. However, anyone who was paying attention could see the direction of travel. Seldom does what begins with support personnel stay there. The agency recruiting The most impacted firms are also reacting in a way that exacerbates rather than resolves the issue.
In the Am Law 200, lateral hiring for lawyers with AI-related experience increased by 68% in 2025, while associate hiring in that field increased by 106% annually. Instead of developing experienced talent, businesses are purchasing it. From quarter to quarter, that makes sense. It silently drains the partnership pipeline for more than ten years. The current associate classes, which are becoming smaller with each cycle, will produce the partners of 2035. LawSites Additionally, there is a persistent client-side squeeze. Nearly two-thirds of internal teams anticipate reducing their reliance on outside counsel, according to a recent survey by the Association of Corporate Counsel and Everlaw.
This is partially due to the fact that they witness firms charging full associate rates for work that a machine can complete in a matter of minutes. Customers are retaliating by directing AI-replaceable work away from outside companies, which directly reduces the income that used to fund sizable junior classes. The reason behind the pyramid’s compression is not the more difficult question. It’s what follows. This is not a scaled-down version of the previous model; rather, it is a smaller base of juniors who have been trained on less volume in an environment where the routine work that developed judgment has been automated.
Most businesses haven’t yet taken that into consideration because it’s a fundamentally different one. In ten years, those who consciously cultivate their smaller cohorts by investing in structured feedback and early substantive responsibility will have a pipeline of partnerships. Those who merely reduce their workforce and believe that the previous training route is still effective are placing a silent wager that is probably going to backfire. The majority of company executives may not even be aware of the decision they are making.
