WASHINGTON, May 28 (Reuters) - U.S. law firm Kirkland & Ellis said it is devoting $500 million of its revenue to developing a custom AI platform, accelerating a spending race on the technology in the legal industry.
Kirkland, with self-reported revenue of $10.6 billion last year, said on Thursday it would invest the funds over the next three to four years, starting with $100 million in 2026.
The firm, founded in Chicago and with thousands of lawyers globally, said it would still license some third-party AI programs. It declined to say if its planned platform would rely on a specific generative AI model.
Major law firms looking to streamline their operations and legal work have emerged as key AI customers, and some have inked deals with leading AI startups to jointly develop legal-focused AI tools. London-based Freshfields said last month it will work with the legal team at Anthropic, which makes Claude, to develop AI applications for legal services.
Kirkland said its platform would be designed based on information from 250 Kirkland lawyers and would involve more than 180 technology professionals inside and outside the firm.
The Financial Times first published details of the plan.
Law firm leaders told Reuters in recent interviews that there is increasing demand for custom-designed AI programs to assist with specific business and legal tasks, and an ongoing debate over how much to develop internally.
Andrew Johnson, chief information officer at law firm Brownstein Hyatt, said five years ago, there was an aversion to the idea of custom development.
“I would say that’s largely not the case anymore,” he said.
Lawyers' increasing AI use carries risks, including data security and the technology's penchant to fabricate case citations, misquote the law or generate non-existent legal sources. U.S. judges have sanctioned lawyers in dozens of cases after attorneys used AI for research or drafting without fully vetting the results.
Wall Street law firm Sullivan & Cromwell apologized to a federal judge last month for submitting a court filing with inaccurate citations and other errors generated by AI.
(Reporting by Mike Scarcella; Editing by David Bario, Rod Nickel)
