It’s claimed that for the first time anywhere in the world, an AI law firm beat a traditionally represented opponent at trial. The implications reach well beyond England
The opposing side had a solicitor and a barrister. Garfield AI had a chatbot and a recently called junior advocate. Last month, at Wandsworth County Court in southwest London, Garfield won.
The case was modest – a £7,000 small claims dispute over unpaid freelance fees – but the result is not. Garfield AI, founded by former City litigator Philip Young and quantum physicist Daniel Long, prepared every document for the three-hour trial, including witness statements, without any human lawyer involvement. A barrister was engaged only for the courtroom advocacy itself. The client, Tamires Camal Taquidir, a freelance HR consultant who had been unable to recover fees from a hospitality business, paid approximately £400 for the service and walked away with a judgment for the full amount. The other side paid for both a solicitor and a barrister and lost.
Philip Young called it "a landmark moment, not just for Garfield AI, but for access to justice." That framing matters less than the arithmetic: a regulated AI firm, charging £2 for a letter and £50 to file a claim, outworked a conventionally staffed legal team in open court.
How Garfield got here
Garfield received approval from the Solicitors Regulation Authority in May 2025 – the first and still only AI-only firm licensed to provide regulated legal services in England and Wales. In the thirteen months since, it has handled more than 600 claims and recovered approximately £500,000 for clients on disputes ranging from £30 to £10,000, almost all settled before any hearing.
The June 22 trial was different. The defendant issued a counterclaim, forcing the matter to court, and Garfield went the distance. SRA chief executive Paul Philip had anticipated something like this when he approved the firm: regulators "cannot afford to pull up the drawbridge on innovations that could have big public benefits." He now has a data point.
The Canadian picture
No equivalent to Garfield AI is licensed to practise in Canada, and the Garfield verdict will not change that immediately. But the profession's preferred framing – that AI is a tool lawyers use, not a service that competes with them – is under pressure.
Canadian firms have spent two years integrating AI into associate-level work: first drafts, document review, research summaries. The effect, as partners at firms including McCarthy Tétrault, Flex Legal Network, and Speigel Nichols Fox have described to Canadian Lawyer, is that output is arriving faster while the senior judgment required to approve it remains the same scarce resource it always was. McCarthy Tétrault's Matthew Peters recently described watching a document review that would have taken six hours complete in one. The bottleneck has simply moved up the org chart.
Accuracy is the more immediate problem. Since a B.C. court flagged Canada's first known case of AI-hallucinated case law in early 2024 – Zhang v. Chen – courts and tribunals across the country have issued over 150 rulings on fabricated citations and invented legal arguments. The Law Society of Ontario has been clear that AI output requires independent human verification; the professional obligations predate the technology and survive it.
The investment keeps coming
The hallucination problem has not slowed the money. In the same week Garfield won its trial, Sullivan & Cromwell disclosed to a U.S. federal bankruptcy court that a recent filing contained multiple AI-generated errors. Weeks before that, UK firm Pinsent Masons was publicly reprimanded by a London court for submitting AI-drafted letters that twice misquoted insolvency legislation. Neither event registered as a pause.
Kirkland & Ellis, the world's highest-grossing law firm, announced last month it is committing $500-million (U.S.) to build a proprietary AI platform. Freshfields signed a deal with Anthropic in April. In Canada, Lavery has launched an internal generative AI tool and Spellbook has put up $1-million in fellowships for law students redesigning legal workflows. The direction is set.
What Garfield's trial win adds to this picture is a proof of concept that the legal profession has not had before: not AI assisting a lawyer, but AI replacing the preparatory work a lawyer would have done, in a regulated context, against a lawyered opponent, with a judge deciding the outcome. The question of when something like that arrives in Canada is now more concrete than it was a week ago.
