TLDR: Jack Newton, CEO of legal tech company Clio, asserts that generative AI tools trained on specific legal data are crucial for accuracy and reliability in the legal profession. Speaking at ClioCon, Newton highlighted the risks of generic AI models, which have led to instances of ‘hallucinated’ case law and lawyer sanctions. He introduced Vincent, Clio’s AI legal assistant, which is grounded in vLex’s extensive legal database, as a solution to provide precise legal information and amplify lawyers’ impact, rather than replace them.
BOSTON, MA – The burgeoning field of generative artificial intelligence (GenAI) is presenting a complex challenge for the legal sector, particularly in distinguishing between reliable legal advice and potentially misleading legal information. Jack Newton, CEO and founder of the British Columbia-based legal tech company Clio, addressed these concerns at the annual ClioCon conference, emphasizing that the efficacy of AI tools in law hinges entirely on the quality and specificity of their training data.
Newton pointed out that while foundational large language models (LLMs) from companies like OpenAI and Anthropic are powerful, their general-purpose nature—being trained on the open web rather than real legal data—makes them prone to inaccuracies. ‘We’ve all seen the headlines of AI making mistakes, hallucinating cases that never existed,’ Newton stated, referencing instances where lawyers have faced sanctions due to reliance on such unreliable AI outputs.
To counter these issues, Clio is championing AI tools specifically trained on legal datasets. Newton announced that Vincent, an AI legal assistant originally launched by vLex in 2018, will now be offered under Clio’s new enterprise division, targeting large law firms and corporate legal departments. Clio’s acquisition of vLex for US$1 billion earlier this summer, marking it as the largest acquisition in legal tech history, underscores the company’s commitment to specialized legal AI.
Vincent’s key differentiator, according to Newton, is its foundation in vLex’s vast legal database, which comprises over one billion legal documents from 180 countries. ‘Unlike generic AI that is trained on data available on the open web, Vincent is legal AI that understands the law,’ Newton explained. He further elaborated that when AI is ‘grounded in legal data, real cases and real decisions, it doesn’t just produce more accurate answers – it gains new capabilities.’
The urgency for such specialized tools is highlighted by recent judicial developments in Canada. Less than two years ago, a BC Supreme Court Justice issued Canada’s first decision concerning the use of AI-hallucinated case law in legal proceedings. Since then, over 30 additional decisions from Canadian courts and tribunals have involved known instances of AI-fabricated content, as tracked by a senior research fellow at HEC Paris.
Also Read:
- Legal AI Market Poised for Substantial Growth, Projected to Reach Billions by 2035
- Chamelio Unveils ‘Agentic Flows’ to Empower In-House Legal Teams with AI-Powered Workflow Automation
Despite some lawyers’ hesitation to adopt AI due to reliability concerns, Newton views the integration of AI capabilities into legal products as a significant opportunity. He projected that it could lead to ‘four times the clients to walk through your doors, four times the matters being opened, four times the revenue flowing into your firm,’ while also addressing the critical ‘access to justice gap.’ Newton concluded by reassuring the audience, ‘AI is not here to replace legal professionals. AI is here to amplify your impact.’
