AI firms promote legal tools as courts penalise fabricated citations in filings
Major AI firms are marketing specialised tools to lawyers even as courts continue to punish lawyers for submitting AI-generated filings containing invented case law. The issue matters because false citations can undermine legal arguments and the justice system, while the industry is presenting AI as a way to support legal work.
A researcher’s database records nearly 2,000 cases of fabricated citations worldwide, involving more than 800 lawyers and 1,100 self-represented litigants. In Canada, reported cases rose from seven in 2024 to 86 in 2025, with 39 more in the first quarter of 2026; one Ontario lawyer was ordered to pay $31,150 after using Grok-generated material. Google and Anthropic say their legal products connect to external legal sources to help ground and verify answers, while the article says SpaceXAI’s legal offering provides less information about how it traces citations to primary sources.
- AI legal products are expanding despite repeated cases of fabricated citations.
- Courts have penalised lawyers for relying on AI-generated false authorities.
- Verification through external legal sources is a key difference between products.
Read the full article at the source →
Originally published by Engadget as “The AI industry wants models to assist in legal battles, but will they help?”.