How Australian law firms can turn AI failures into success

Lawyers have had a few years to wrestle with AI and integrate these changes into their practices, but the conversation has never been more urgent – this is the moment to transform early missteps into game-changing lessons and tangible progress.

Speaking at Lawyers Weekly AI Innovate webcast, three leading voices from Thomson Reuters exposed the biggest AI blunders they’ve seen and revealed how lawyers can turn these failures into powerful lessons to make 2026 a year of smarter, faster, and more confident AI-powered practice.

The cautionary tales shaping the narrative

Across the legal landscape, a string of high-profile AI missteps has sent shockwaves through law firms and courts alike, casting a long shadow over the adoption of AI in legal practice.

Fiona McLay, lead AI client adviser at Thomson Reuters, pointed out that these high-profile cases share a common risk: courts receiving documents riddled with fabricated case citations.

“There have been lots of headline cases, and there seems to be a common theme to the cases we’re seeing,” McLay said.

VIEW ALL

“Where somebody has prepared documents that have been filed with the court, and they have included false citations of cases which don’t exist or have been cited as authority for a proposition where they are not.”

Most striking, McLay warned, is that despite repeated guidance from law societies, practitioners continue to claim ignorance – many lawyers still don’t realise AI tools like ChatGPT should never generate unverified submissions.

“I spend a lot of time talking to lawyers about AI and ways it can be used to help, so I’m a bit surprised that people are still saying to the judge, I didn’t know that I couldn’t use ChatGPT to write submissions,” McLay said.

“Certainly, the law associations, lots of bodies have been trying to actively pass that message to the profession.”

The wrong message is being sent

The headline-grabbing AI failures have sent a blunt message through the legal profession, with some believing that AI must be off-limits and should not be used by lawyers at all.

However, McLay warned that this widespread perception is misleading and emphasised that the real takeaway for the profession is that AI can be used safely while ensuring no unchecked words are placed into court documents.

“What I am seeing is a bit of a side or a flow-on effect from some of these headlines is people say, ‘Oh, we can’t use AI, it’s not safe,’” McLay said.

“I don’t think that that’s the right message either. I think there’s definitely a safe path. I think the message is you shouldn’t put words in court documents that you haven’t checked.”

The real danger, she warned, lies not in AI itself, but in choosing tools that make verification difficult, leaving lawyers unable to check sources with the care demanded of professional work.

“I think it is important to look for tools where it’s easy to have that check and have that verification and the source be right there in the same screen,” McLay said.

“Choosing a tool where you can be working in Word and do your verification in an interface that lawyers are comfortable with, I think that’s going to always give you better adoption.”

From mistakes to mastery

While the lessons emerging from headline AI failures are undeniably valuable, their true impact will depend on whether lawyers are prepared not just to take them in, but to meaningfully act on them.

Jen Lee, product strategy director at Thomson Reuters, struck an optimistic note on the profession’s outlook, observing that while some lawyers remain hesitant, the broader sentiment is overwhelmingly positive.

“Look, I would say I’m very optimistic. Perhaps it’s the nature of my job, but generally, the people that I speak to are quite enthusiastic about the tech,” Lee said.

“I do also speak to people who have reservations, but actually, those are the conversations that I like because I think it’s much more interesting to have that conversation with someone who is averse to AI and to be able to talk through it and then to potentially see a shift.

“But overall, I’d say very enthusiastic, and I don’t think lawyers and their roles are going anywhere soon. It’s just that nature will change somewhat.”

While Ziggy Cheng, legal AI specialist at Thomson Reuters, echoed this optimism, he cautioned that the path forward will be far from smooth, with bursts of enthusiasm often giving way to hesitation as some lawyers pull back amid lingering concerns around its use.

“I’m very optimistic. But it’s going to go up and down. It’s going to be a bit of a roller coaster,” Cheng said.

“A lot of people actually were just scared of using the AI, and so we’ve had to pull that back a little bit.”

While noting that high-profile missteps will keep surfacing and momentarily shake confidence, Cheng stressed that the underlying momentum is unstoppable, and lawyers will ultimately have to embrace AI and master its responsible use.

“But these kinds of events will continue to happen. There will be some big profile cases of people just really using AI in the wrong way and getting caught out for it,” Cheng said.

“It will shock people, and they’re going to put the tools down a little bit, but the momentum is strong. The fact of the matter is this isn’t going away, and people are going to need to learn this and use it properly.”

Winning with AI in 2026

To turn these cautionary tales into clear lessons, law firms of all sizes can take steps not only to avoid the pitfalls of headline-making AI failures but also to seize the opportunity to redefine how legal work is done.

While many lawyers later in their careers may feel they don’t need a plan for AI, Cheng warned that avoiding it is no shortcut and that the rest of their career will feel much longer if they don’t get on board.

“Start simple. I mean, you have to get the tools. I speak to some people who say, I’m only going to be practising for another five years, and so I’m just not going to bother learning this technology,” Cheng said.

“Well, that’s going to be a long five years for you if you don’t jump on board.”

Cheng emphasised that adopting AI doesn’t have to be overwhelming, urging lawyers to start small, master a single use case, and embrace continuous learning to steadily build confidence and skill over time.

“It’s not that hard. You don’t have to learn everything. Start with one use case, get to know that, understand it, maybe share it with your colleagues, teach them how to do it, and then they can teach you some stuff back,” Cheng said.

“Start small, but you have to get in there and start using it and of course, the continuous learning.”

To stay ahead with this evolving technology, McLay advised that the most successful AI strategy for lawyers begins with curiosity, exploring what the tools can do, and focusing on a few key use cases before expanding from there.

“I think it’s being open to being curious, getting a level of awareness about what these tools can do and then focusing in on a couple of key use cases,” McLay said.

“I think from using that as a bit of a foundation and then building on from there, that’s the strategy I’ve seen be most successful.”

Lee urged lawyers not to work in isolation, stressing that staying ahead means understanding both what your firm is doing and what your clients truly expect and need from you.

“I would say not doing things in isolation. Our clients are all using AI, and so I think it’s really important to be aware of not just what your firm organisation is doing, but what your clients are doing and what they expect in terms of transparency and communication,” Lee said.