October 9, 2026, (Inside AI) — Big Law is starting to pay its lawyers to learn AI. Ropes & Gray, a Boston-founded firm with more than 1,000 associates and counsel, now lets those attorneys count up to 100 hours of AI experimentation and training toward annual billing requirements. The policy expands a 2025 pilot that gave first-year associates credit for nearly 400 hours of AI work, according to a firm spokesperson.
The move matters because billable hours remain the core productivity metric in private practice. They determine bonuses, staffing, and promotion. When a firm carves out paid time for AI training, it is effectively betting that short-term billable losses will be offset by long-term efficiency gains. That bet is spreading across the industry.
At Akerman, lawyers have been able to earn credit since April for AI projects that "really contributing value" to the firm, chairman Scott Meyers said. The firm runs an internal portal where associates can pick up AI assignments or pitch their own. Projects so far include an AI agent that reviews construction contracts and tests of AI research tools to document where they produce inaccurate results.
"We understand that there may be some compression in billable hours," Meyers said. "If we need to find other ways to fill that time, I suspect it will be that combination of billable and creditable hours."
Read: AI Reshapes Tasks, Not Jobs, ADP Economist Says
That compression is already visible elsewhere. AT&T's legal department has stopped paying outside firms for many routine assignments once handled by junior lawyers. It now uses an in-house team powered by AI, and the company's general counsel said its roster of roughly 30 outside firms could shrink further.
Ropes & Gray's program is not just a perk. It is part of a firmwide training initiative that has been running since the spring. Lawyers move through modules toward applying AI to client work, then are encouraged to build internal AI tools or processes, said Patrick Ryan, the firm's director of knowledge management innovation. He called the first-year program a "huge success" and said all of the firm's first-year associates participated.
A Ropes spokesperson said the firm has "invested over 35,000 associate hours to AI time this year," but did not specify whether all of those hours were covered by the credit program.
For associates, the credit removes a practical barrier. Learning to prompt AI tools well takes time, and that time does not bill to a client. Kevin Schoonveld, a New York-based ninth-year debt finance associate at Ropes, said the learning curve is real.
"When you're first playing around with these AI tools, you're investing a lot of time in it and you're not seeing a ton of efficiencies at first," he said.
Two Ropes associates, who spoke on condition of anonymity, said the credit gives them room to test tools even when the work cannot be billed directly. That flexibility may prove essential as firms push AI into client matters. It also raises a question firms have mostly avoided in public: if AI handles more work, what happens to the associate pyramid that has long fed the partnership track?
Read: Korn Ferry: There Is No Such Thing as an AI-Ready Culture
Meyers said compression alone should not dictate fewer associates. He argued the credit program helps address associates' concerns about AI's impact on their billable time. Most large firms already grant billable credits for pro bono work, so the structure is familiar. The difference is that AI credits pay lawyers to improve the firm's own technology and workflows, not to serve clients or the public.
Ropes is also folding AI into annual review conversations, but not to penalize lawyers. Amy Ross, the firm's chief legal talent officer, said the goal is to spot gaps.
"To the extent that we're finding people aren't using it, we want to address that too," she said.
Outside the billable-hours experiment, the legal industry's AI transition is producing mixed signals. AI-generated court filings riddled with errors continue to surge despite three years of sanctions, and law firms are competing hard for technology talent. Meanwhile, Kirkland & Ellis, the highest-grossing US law firm, told The American Lawyer it will stop reporting annual financial results. Kirkland said public reporting of revenue and profits "does not provide meaningful value to our clients nor adequately reflect the caliber of our firm's legal services." The firm's revenue topped $10.56 billion in 2025, with average profits per equity partner of $11.1 million.
For now, the credit programs at Ropes and Akerman offer a rare concrete answer to a question that has lingered since generative AI reached the legal sector: who pays for the learning? The firms are. Whether that investment produces durable advantages, or simply delays a harder reckoning over leverage and staffing, will depend on how quickly AI moves from experiment to billable work.
