Law firms probe best ways to win a head start on using AI

Danny Gilbert, veteran corporate lawyer and founder of Australia’s Gilbert + Tobin, notes that profound changes are looming for the legal sector — and he is keen to spell out why.

“Generative AI is coming at us; it will gather and analyse all the data you need in major corporate deals,” he says. “It will predict what will happen in court, examine the submissions of opposing counsel and even produce a good first draft of a judgment.”

In the 18 months since the launch of OpenAI’s ChatGPT, many law firms in the Asia-Pacific region have moved quickly to embrace this new technology — and even try to get a head start.

Paul Jenkins, global chair of international law firm Ashurst, says: “Artificial intelligence is a significant change for us — to ensure that we not only keep up, but we stay ahead.” Ashurst has conducted pilots and trials involving more than 400 staff, in 23 offices, in 15 countries. “One of the trials showed we can save an average of 45 per cent of [the] time — and, in the most positive case, up to 80 per cent of [the] time — on preparing the first draft of client briefings.”

The efforts of law firms in the region to stay on top of the developing technology show the profession is focused on adapting. From legal documents that can be produced in a minute with 80 per cent accuracy, via MinterEllison, to human holograms that can talk to potential clients, deployed by Corrs Chambers Westgarth, the speed of AI experimentation is striking.

At a recent global round-table involving 25 legal industry leaders, hosted by the Financial Times, there was a consensus that the law firm of the future would be unrecognisable to the current generation. For a start, many firms have already moved into consulting services in areas adjacent to the law, such as advising on general business risk and implementing regulatory compliance programmes. At Ashurst’s consulting business, for example, annual income grew by just over 60 per cent last year.

Gilbert observes that a multidisciplinary approach could become the norm, and that even firms that want to stick to their legal knitting will hire many more scientists and technologists to deliver legal advice.

So far, law firms in the Asia-Pacific region are using generative AI for tasks ranging from research to creating, analysing and summarising legal documents. Some firms, not just the biggest, have set up proprietorial AI assistants — including Clifford Chance Assist, KWM Chat and Ask. KAI from Khaitan & Co. The firms claim the benefits are immediate. In a King & Wood Mallesons survey of 900-plus active users, 18 per cent said KWM Chat saves between five and 10 hours a week, while more than 70 per cent said it saves up to four hours.

Different foundation models [of generative AI] show materially different performance and cost. Where the balance lies . . . is what we need to work out

But many firms also sound a note of caution about generative AI’s immediate potential. Ben Allgrove, a partner and seasoned legal technologist at Baker McKenzie, the global law firm, warns that “generative AI hype has echoes of 2017”, when much-heralded advances in machine learning turned out to be modest.

It does mark a significant step up from previous iterations of artificial intelligence, he says, but Baker McKenzie is focusing its experimentation on discrete, client-specific tasks where the data is known to be solid before making big bets on particular AI models.

Language informs culture and vice versa, and I hope this AI will bring a better understanding of people

For Allgrove, one of the most important questions is simply the outlay. Rolling out Microsoft Copilot, Harvey or other generative AI tools to 4,700 fee-earners, let alone all 13,000 staff, looks too expensive when the return on investment remains uncertain. “Different foundation models show materially different performance and cost. Where the balance lies between them is what we need to work out in years to come,” he says.

In addition, there are language issues. Ryutaro Nakayama, managing partner of Nishimura & Asahi, describes the challenge for his firm: “The Japanese language is more complicated than English. We have fewer available data sets, and a lot of Japanese contracts are still paper.” Instead, he envisages the biggest changes at Japanese law firms in the near future will come via a convergence of professional service advisers, where law firms offer adjacent consulting services.

Across the region, though, the short-term productivity gains from using generative AI — even for everyday tasks — look appreciable. “The jump could be better than [that generated by] email in the 1990s,” says Allgrove. “But there is a strategic question as to where those gains will land and where to jump.”

For now, law firm leaders in the region are upbeat about AI developments. Sue Kench, global chief executive at KWM, observes that, while change will arrive slower than some predict, she is already noting benefits to organisational culture at the Australian-Chinese firm.

“I was in one of our regular China management meetings and the improved WeChat [text] translation was better than having a [human] translation in my ear,” she says. “Language informs culture and vice versa, and I hope this AI will bring a better understanding of people.”