How should you react to a metal cart striking your leg during a long-haul flight into New York’s John F. Kennedy International Airport? Well, after muttering, “Oh dear, that rather smarts,” you sue the airline – which is what Roberto Mata did after alleging that just such a cart struck his left knee on an Avianca flight from Costa Rica. After some legal to-ing and fro-ing, Mata’s lawyers submitted a brief to the court stuffed with precedents supporting their client’s argument for redress. The problem was that none of these cases had ever taken place. Mata’s lawyer, Avianca’s counsel discovered, had used ChatGPT to write his brief.
The lawyer in question did not escape a scolding from the presiding judge. “The opposing party waste[d] time and money in exposing the deception,” he wrote, not to mention depriving themselves of the opportunity to strengthen their case by using real case law. Above all, Judge Castel continued, using AI in this way “promotes cynicism about the legal profession and the American judicial system.”
A one-off? Well, no. A year later in Canada, a Vancouver lawyer relied on fictitious cases generated by OpenAI’s chatbot in a family law case. Then, 8,200m away, a solicitor using AI-enabled legal software in Victoria, Australia, handed a list of what purported to be prior cases requested by the judge in another family dispute, none of which could be corroborated in the official record. Earlier this year, meanwhile, a UK barrister was accused of misleading an immigration tribunal by citing a fictitious Court of Appeal judgment generated by ChatGPT.
These are by no means the only examples of the legal industry’s ineptitude in using AI, though some of their clients are having more success. In the last month, a woman in California successfully overturned an eviction notice and avoided hefty financial penalties by using AI tools, including ChatGPT, to mount her own appeal. Some lawyers may have been humbled by their ineptitude with AI, but nothing seems likely to deter the industry from giving it more and more responsibility in the legal process. After all, AI takes out the drudgery of routine tasks and lets them focus on complex issues of law and the building of client relationships.
It’s for these reasons that the industry has relied on AI for almost a decade. “Legal is usually lagging when it comes to technology adoption, but it was in the vanguard when it came to GenAI,” says Joanna Goodman, author of ‘Robots in Law: How Artificial Intelligence is Transforming Legal Services’, which, when it was published in 2015, was one of the first books written about legal AI. Since then, pioneers like Luminance, RAVN, and Kira enabled information retrieval, document comparison and analysis, and more.
This isn’t just AI froth. Across the board, law firms have started to buy heavily into the idea that AI can expedite much of the dogsbody work that has come to define the back office life of your average lawyer. A 2024 survey of practices by LexisNexis, which provides data and analytics to legal professionals, noted a profound shift taking place, with the number of lawyers with no plans to adopt AI dropping sharply from 61% to just 15%. Moreover, the number using GenAI for work purposes had almost quadrupled in little over a year, from 11% to 41%.
The risk of falling foul of ChatGPT’s habit of ‘hallucinating’ false but convincing information looms large over, particularly in a profession steeped in history, heritage, and tradition – barristers’ wigs and gowns hardly scream modernisation. But as the industry puts its faith in GenAI and its unparalleled ability to read, summarise and generate text, how much machine assistance is too much?
Adapt or die
Perhaps surprisingly, law is among the leading AI adopters. According to the 2025 Generative AI in Professional Services Report from Thomson Reuters, law firms and corporate legal departments are the top two professional practices already using GenAI. But its impact has not been uniform across the legal industry. According to the Solicitor’s Regulation Authority (SRA), around 75% of the UK’s largest law firms are using AI. In contrast, only 30% of small law firms are implementing it.
Larger firms may have the internal infrastructure to lead the charge, so smaller firms may be left behind. AI is undoubtedly on the agenda for every law firm CEO or managing partner, but those stuck in the early stages of implementation are at a disadvantage.
This is less of a problem for bigger firms, where AI has already become indispensable for automating repetitive tasks like contract analysis, e-discovery and deal due diligence. According to Goodman, most of these practices apply a build-and-buy GenAI strategy, with the emergence of legal-specific GenAI platforms like Harvey and Legora expediting legal tasks and processes.
Some of the biggest firms have reportedly even started building their own GenAI chatbots using GPT platforms from OpenAI and other providers. Linklaters has developed its own GenAI chatbot named Laila, using Microsoft’s Azure OpenAI infrastructure. A&O Shearman Microsoft Copilot is another easy option, as Microsoft 365 is widely used, and the technology can operate behind the firewall to avoid the data security and confidentiality issues with public platforms. Indeed, Clifford Chance has used it as the basis for its own secured AI tool – Clifford Chance Assist.
Bird & Bird is an international firm known for its tech and innovation focus, particularly in areas like data privacy, IT, and telecoms. Its lawyers are extensively leveraging AI to streamline legal processes, reduce lawyers’ non-billable hours, enhance accuracy, and increase productivity. Its head of legal tech and innovation, Hélder Santos, notes that a contract analysis that once took days can now be completed in under an hour. “This helps deliver faster results for our clients and enables our lawyers to use the time saved to provide tailored solutions like scenario-specific playbooks, adding that extra expert layer of human insight,” he says.
He estimates that three-quarters of the UK’s largest firms have a GenAI toolkit, and globally, the LegalTech market is poised to surpass $20bn within a few years, noting that “AI has evolved from a curiosity to an expectation, reshaping how legal work is conducted every day.”
The legal market will not evolve uniformly, however, and different models will no doubt emerge. Some firms will build and own their own platforms, while others will leverage AI to deliver niche services at scale and speed. Some experts, including Santos, believe that multidisciplinary teams blending lawyers, technologists, data experts, and strategists will become the norm.
“At the same time, however, some firms will cling to tradition, perfecting the old model while hoping the new world will pass them by,” he says. “But history has shown what happens to those who resist technological change. They risk becoming master craftsmen of horse carriages, polishing their work while automobiles speed past.”
Rise of the AI-powered lawyer
Some clients stand to benefit greatly from faster or better advice, as AI tools help legal teams leverage their experience over multiple matters to create bespoke playbooks for specific scenarios. And AI could change the legal pricing model. Law firms charge a hefty fee per lawyer-hour – with GenAI capable of performing some tasks instantly, the justification for that pricing framework suddenly looks shaky at best.
Clients of larger firms that are willing to invest in the technology and the related skills and training will see the greatest benefit, argues Rachel Broquard, but AI is already delivering value for smaller clients in areas like M&A and litigation support. “Even with real estate, employment advice or routine NDA work,” says Evershed Sutherland’s service excellence partner, “we are seeing clients benefit and making legal advice more accessible to all.”
GenAI can free up lawyers to focus on higher value strategic work, potentially delivering faster, more accurate and more efficient outcomes. It could fundamentally change their role, which has been defined for generations by three pillars: legal expertise, client relationships, and financial contribution. For Santos, those pillars remain essential, but the partner of the future must be something more – an architect of innovation.
“That means seeing law not only as a profession but as a system that can be redesigned, packaging knowledge so it can be reused at scale, integrating legal services directly into client processes, and shaping business models that go beyond the billable hour,” he says. “In short, it requires a shift from delivering services to orchestrating how services are created and experienced.”
As well as changing how lawyers work, effective digital systems and automation could revolutionise consumer law and court processes, and – with the right level of investment – may help people understand and exercise their rights. Nevertheless, the jury is out on how well GenAI contributes to decision-making. While the technology can generate answers at speed and scale, it cannot guarantee accuracy, context, or relevance, as those lawyers who relied too heavily on ChatGPT found out to their cost. Ironically, fee-earners will need to use some of the time GenAI saves to thoroughly review and validate its output.
“In a profession built on trust and human judgment, accountability cannot be delegated,” stresses Santos. “Leading with strong governance and ethics isn’t just good practice – it’s a marker of business maturity and responsibility in a rapidly evolving landscape.”
When it comes to oversight, a survey by the International Bar Association (IBA) and the Centre for AI and Digital Policy (CAIDP) in 2024 raises some concerns. It found that larger law firms are using AI in client-facing applications, but less than half of the respondents (91 out of 210 firms) had policies in place around its use within their organisations.
Consequently, lawyers will need to not only develop prompting skills to get the best outcomes, but also ensure that they are only using enterprise-approved tools to maintain client confidentiality and data security. “They also need to adjust their approach to supervision and apply critical thinking and their expertise to ensure that AI-generated outputs are accurate and contextualised,” says Broquard.
The lack of internal expertise and concerns around data security and regulatory compliance could stem the tide of GenAI, but not for long. Indeed, Broquard points out that many clients were cautious about the use of their data in law firm AI tools early on in the GenAI journey, but they are now more comfortable that firms are implementing AI responsibly.
As agentic AI starts to make inroads into the sector and AI-enhanced workflows are used more extensively, trust and ethical oversight will be critical, as will technological proficiency. Any lawyers who simply ask ChatGPT to build a case will soon be wondering why they wasted so many years in law school just to end up flipping burgers.
Read more: Agentic AI needs stronger digital certificates
