Artificial intelligence adoption in the legal profession is accelerating rapidly, but so are concerns around accuracy, accountability, and trust.
A 2025 survey found that 80% of legal professionals reported using AI last year, up from just 22% in 2024. As usage grows, so does the risk of relying on tools that can generate incorrect or unverifiable information—something lawyers cannot afford.
“The consequences of errors are real. Someone’s freedom, reputation or livelihood may be on the line,” said Rawia Ashraf, Head of Product, CoCounsel Legal at Thomson Reuters. “That’s why the AI lawyers use must be held to the same professional standard they are. It can’t cut corners. It has to be grounded in trusted law, show its work, and deliver answers that stand up in court and in practice.”
The shift is pushing legal technology providers to emphasize trust, transparency, and verifiable outputs as core requirements. Companies like Thomson Reuters are positioning their platforms around what they describe as “fiduciary-grade AI,” designed to protect customer data while delivering research and analysis that can withstand professional scrutiny.
Canadian legal teams, like Hull & Hull LLP, are already beginning to apply these tools in practice. Features such as Deep Research on Westlaw Advantage allow lawyers to submit plain-language queries and receive fully sourced research summaries in minutes, reducing the time required for traditional legal research.
Thomson Reuters is also bringing its CoCounsel Legal platform to Canada, an agentic AI system designed to complete complex legal workflows from a single query by planning and executing multiple steps using authoritative legal content. Globally, the company says more than one million professionals are already using CoCounsel across 107 countries and territories.
Even as capabilities expand, the role of human oversight remains central.
“This isn’t a replacement for legal work or human judgment,” said Ashraf. “Legal professionals are in the driver’s seat. It’s very important for humans to be in the loop in this work.”
As AI becomes more embedded in legal workflows, the defining question is shifting from whether firms will adopt these tools to how much they can trust them.