When US lawyers in an aviation injury claim admitted that six cases they had cited in a legal brief against Colombian airline Avianca in 2023 had been entirely invented by ChatGPT, there were immediate concerns about using generative AI in a legal setting.
Since then, numerous dedicated legal GenAI platforms have emerged that are trained on legal data, reducing the risk of these ‘hallucinations’. This has helped assuage fears, with GenAI usage among lawyers jumping significantly over the past 12 months. Some 26% of legal respondents in Thomson Reuters Institute’s 2025 Generative AI in Professional Services Report say their law firms are using GenAI on an enterprise level, up from 14% just a year ago.
“The reason you are seeing that jump is just because it takes a little bit of time, especially for law firms that are a little bit more risk averse by nature,” says Zachary Warren, technology and innovation insights lead at Thomson Reuters Institute.
While 2023 was a year of education, with firms trying to understand what AI could do, and 2024 was about trialling the tech and running pilot projects with small user groups, this year has seen more widespread adoption of GenAI tech among legal professionals.
“I would expect another jump in 2026 as more and more firms go through that progression,” says Warren.
Another driver for the jump in usage is client demand. Some 59% of corporate legal departments said they want their law firms to use GenAI, with just 13% opposed.
“There’s a little bit more pressure on law firms to abide by what their clients want, and even in some cases if firms think they’re not ready, their clients are making them ready to take this leap,” says Warren.
While law firms may have struggled in the past with getting buy-in for new technology projects, GenAI’s ubiquity in everyday life means many legal professionals are much more comfortable about embracing the technology.
“I have never seen our lawyers as excited about technology,” says Rachel Broquard, service excellence partner at Eversheds Sutherland. “The challenge is there’s been an explosion of new technologies coming to the legal market, so you have to have a strategy that aligns to the firm strategy and you’ve got to stick to it.”
Rather than struggling to convince lawyers to adopt GenAI, firms face the opposite problem: setting boundaries on which GenAI tools can be used and how.
“The challenge around buy-in is more about getting everybody to understand why it is that we’re doing what we’re doing, and why we’re saying no to everything else,” says Broquard.
Even though usage levels are increasing, GenAI’s impact on the profession is still fairly limited, in part because the adoption is new, but also because firms are unsure how to measure return on investment (ROI). Only 20% of respondents in the Thomson Reuters Institute GenAI survey said they are measuring ROI.
Of that 20%, some are doing this through internal measurements such as cost savings (79%), usage (64%) and satisfaction (51%). However, a smaller cohort of those measuring ROI are looking at client satisfaction (38%) and where AI has helped win new business (24%).
“That’s not where firms are now, but it’s probably where they need to be going in the near future,” says Warren.
The adoption of GenAI is also likely to change client expectations around service delivery, with a greater focus on how firms can add value in a world where corporate legal departments potentially have access to the same AI tools.
Firms will also have to rethink how to train new lawyers given much of the work junior lawyers have traditionally done will be replaced by AI.
“We are 100% going to have to train people differently because we’re going to work differently,” says Hunter Jackson, chief knowledge officer at McDermott Will & Emery. “People have to be interested in this and roll up their sleeves and learn it, or they risk being left behind by those people who do.”
People have to be interested in this and roll up their sleeves and learn it, or they risk being left behind by those people who do
This is leading to worries about job security. Some 44% of respondents in Thomson Reuters Institute’s survey said they think GenAI is ‘somewhat of a threat’ to jobs, with 15% saying it is a major threat.
“The most frequent question that I get from any junior associate is, is this going to put me out of a job?” says Jackson. “My answer to that is, it’s not going to put you out of a job, but somebody who understands this technology very well might.”
While there is still hesitancy around the future of GenAI, this is dissipating, with the number of respondents in the Thomson Reuters Institute survey with feelings of hesitancy falling to 24% this year from 35% in 2024.
“Once you become hands-on with how GenAI works and you have realistic expectations about how to use these tools, that hesitancy goes away,” says Warren.
The potential for GenAI is also expected to grow as the tech becomes more advanced.
“GenAI is the worst it’s ever going to be – it’s only going to keep getting better from here on in,” says Broquard. “People are excited about it. People are curious. And curiosity is a really important skill around all of this. But ultimately it’s not about the technology, it’s about what you do with it to provide great service.”
Some legal professionals believe GenAI is at an inflection point that will fundamentally transform the business of law.
“You can’t stop the change coming, so we can either decide whether we want to be at the forefront and cutting edge of innovation or look up in four or five years and say ‘wow, we bled 20% of our deal work to XYZ firm because they adopted this technology and were able to work better and smarter and faster’,” says Jackson. “We don’t want to be that organisation.”
The pace of change is likely to pick up considerably. While only 15% of law firm respondents say GenAI is central to their workflow now, that is expected to rise to 95% in the next five years. If the jury is still out on GenAI in the legal sector, this would suggest those deliberations will be coming to an end very soon.
To learn more about AI you can trust, visit: www.thomsonreuters.com
When US lawyers in an aviation injury claim admitted that six cases they had cited in a legal brief against Colombian airline Avianca in 2023 had been entirely invented by ChatGPT, there were immediate concerns about using generative AI in a legal setting.
Since then, numerous dedicated legal GenAI platforms have emerged that are trained on legal data, reducing the risk of these ‘hallucinations’. This has helped assuage fears, with GenAI usage among lawyers jumping significantly over the past 12 months. Some 26% of legal respondents in Thomson Reuters Institute’s 2025 Generative AI in Professional Services Report say their law firms are using GenAI on an enterprise level, up from 14% just a year ago.
“The reason you are seeing that jump is just because it takes a little bit of time, especially for law firms that are a little bit more risk averse by nature,” says Zachary Warren, technology and innovation insights lead at Thomson Reuters Institute.
