Legal AI has a credibility problem.
Courts, lawyers, and the public are increasingly relying on AI tools to answer legal questions, yet many of these systems are trained on scraped internet content rather than official court decisions. The result has been hallucinated citations, fabricated precedents, and growing concern about whether AI can be trusted in high-stakes legal settings.
In response, the University of British Columbia and Vancouver-based startup Caseway have launched a two-year research collaboration to build a legal AI system grounded exclusively in real court decisions. The project aims to dramatically reduce hallucinations and improve the reliability of legal research AI in Canada.
As General-Purpose AI Pulls Back from Legal Use
The timing of this initiative is significant. In November 2025, OpenAI updated ChatGPT’s usage policies to prohibit the AI from providing legal advice, a change that has stirred debate across the technology industry. This retrenchment by the world’s most popular AI chatbot has raised concerns about a new gap in legal help for the public. Many individuals, unable to afford a lawyer who might charge $400/hour, had begun turning to AI for guidance on legal questions.
If general AI models step back from legal assistance, who will fill that void? The UBC–Caseway team believes domain-specific AI is the answer. Building a custom LLM-based legal software is viewed as necessary to overcome the limitations of general-purpose models like ChatGPT, including their tendency toward errors and misinformation. By focusing narrowly on legal knowledge and reasoning, the new model aims to provide dependable answers that ordinary people can use with greater confidence.
Caseway’s database already indexes millions of court decisions
“Access to justice is a huge problem, here in Canada, roughly 70% of people in the justice system lack legal representation,” said Caseway’s Alistair Vigier. “With ChatGPT pulling back, that’s one less tool people might have experimented with for help. We see our legal AI as a way to level the playing field. Whether you’re a lawyer at a law firm or someone preparing for court on your own, this technology will offer reliable legal research at your fingertips, so you’re not left in the dark.”
The need is indeed pressing: a majority of those navigating civil and family courts on their own often struggle to interpret laws and find relevant precedents.
Vigier says that if everyday Canadians could ask an AI trained on official court decisions and get answers without the usual AI fabrications, it could empower them to understand their rights and options better. He also points out that misinterpreting the law or missing a key case can have life-changing consequences.
“Getting the right information quickly can make all the difference. We’re not building this to replace lawyers, but to make legal knowledge far more accessible to those who can’t easily get professional help.
The problem
Every day, thousands of Canadians appear in court without a lawyer. Many turn to online searches for help, and find misinformation instead. The Caseway-UBC partnership project aims to change that.
The project’s goal is to dramatically reduce AI “hallucinations,” instances where generative AI produces false or made-up information, and improve the accuracy of legal research software like Caseway. By combining UBC’s cutting-edge natural language processing research with Caseway’s real-world legal data and expertise, the team plans to create one of the first dedicated legal artificial intelligence models in the world. This home-grown Canadian initiative comes at a pivotal moment, as mainstream AI systems grapple with reliability issues and new restrictions on giving legal advice.
Across the world, governments and researchers are wrestling with how to make AI more accountable. In fields like law, where precision can determine someone’s freedom, financial security, or access to justice, even one false citation can have real-world consequences.
Two-Year NSERC and Mitacs Project Tackles AI Hallucinations in Law
The partnership brings together Dr. Vered Shwartz, an Assistant Professor of Computer Science at UBC and Canada CIFAR AI Chair, and Alistair Vigier, the CEO and co-founder of Caseway. Backed by research grants from NSERC and Mitacs, their joint project will span two years and support a postdoc and a graduate student to develop an accurate and reliable LLM-based legal research software.
Uniquely, the tool will be built on open-source AI foundations, allowing the team to customize it for legal applications while ensuring transparency and potential for wider community use.
“This partnership is a game-changer for legal technology,” said Alistair Vigier, CEO of Caseway. “Combining Caseway’s practical innovation with UBC’s world-class AI research, we’re poised to create something truly transformative. We want to make sure anyone, from lawyers to self-represented people, can get reliable legal answers when they need them.”
Natural language processing and responsible AI
Dr. Shwartz’s lab specializes in natural language processing and responsible AI applications in sensitive domains like law. One of the central challenges the team will tackle is the “hallucination” problem in current AI. General-purpose models like ChatGPT have been prone to confidently output incorrect legal citations or non-existent precedents, which can mislead users. These mistakes by AI pose obvious risks in law, where accuracy and truth are paramount.
“LLMs are already being used in fields such as law, but they still suffer from limitations like overconfidence and the tendency to hallucinate facts,” Dr. Vered Shwartz noted, citing her concern that rushing AI into legal use without fixes could be harmful.
“People are already using LLMs to get answers for legal questions. Our goal is to deliver these answers in the context of Canadian law, with source attribution and reducing hallucinations. By focusing on tasks such as improving the granularity of the retrieval of case laws, combining LLMs with symbolic reasoning, and rigorously testing the tool, we aim to create an AI that professionals and the public can trust to provide accurate legal information,” said Shwartz.
Towards Affordable and Inclusive Legal Research
Both UBC and Caseway view this collaboration as an opportunity to advance access to justice, particularly for marginalized communities. Caseway has also been in discussions with groups such as the Indigenous Bar Association on how AI software could help address unique legal information needs in Indigenous communities.
Cultural and linguistic nuances, which Dr. Shwartz’s research also emphasizes, will be considered in training the model, so that it can serve diverse user groups responsibly. “We want to ensure this technology benefits all Canadians, including those who have historically faced barriers in the legal system,” Vigier said.
In line with that mission, Caseway is planning a shift to a freemium model for its services in the future. While today a subscription is required, the company is actively seeking grants and public funding to offer core legal research features for free to the public.
Caseway notably avoids traditional venture capital and instead leverages angel investors and Canadian government grants to maintain an access-focused approach. The development of a free, AI legal research platform is already underway, which would allow anyone to pose legal questions and receive answers backed by actual case law at no cost. “Our vision is to provide the average person the same research power that lawyers have, without the hefty price tag,” said Vigier, noting that sustainable grant support will enable the company to keep the lights on while offering a basic version of the software to all.
A Breakthrough for Legal Tech and Academia
This UBC–Caseway partnership highlights how academia and industry can join forces to solve real-world problems. It also positions Canada at the forefront of legal AI innovation.
By building a specialized legal research tool that prioritizes factual accuracy and accountability, the team hopes to set a new standard for AI in the legal field. Media watchers have noted that the project’s objectives align with growing calls for responsible AI development, creating systems that are not only powerful, but also transparent and safe for public use.
If successful, the legal AI developed here could become a blueprint for similar efforts in other countries and domains. For now, the collaboration is just beginning, with research work kicking off this fall at UBC’s Vancouver campus and Caseway’s data engineers working in tandem. All eyes in the legal tech community will be on this initiative over the next two years.
How legal research is conducted
A successful outcome could fundamentally change how legal research is conducted, slash the time and cost required to find answers, and open up the justice system to millions more. As major AI companies retreat from high-stakes domains like law and medicine, Canadian innovators are stepping up to build domain-specific systems that prioritize accuracy over scale.
If successful, this initiative will not only transform legal research but also demonstrate that Canada can lead the world in responsible AI. “We’re building a model that reflects Canadian values of fairness, transparency, and inclusion,” said Alistair Vigier. “The law should be understandable to everyone, and this technology brings that goal within reach.”
