Anthropic Lawyer Apologizes After Claude AI Hallucinates Legal Citation in Copyright Lawsuit

Anthropic's Claude AI chatbot generated a false legal citation during the company's ongoing copyright lawsuit with music publishers, leading to a public apology from Anthropic and its lawyers. The erroneous citation, featuring an inaccurate article title and incorrect authors, slipped through manual citation checks and appeared in legal filings and expert testimony. The incident highlights the risks of AI hallucinations in high-stakes legal environments and has drawn scrutiny from a federal judge, Susan van Keulen, who ordered Anthropic to address these issues. Despite the mistake, Anthropic asserted the error was unintentional and not an attempt to fabricate authority, emphasizing that the cited article genuinely exists. This case contributes to broader concerns about AI reliability in legal practice, as other firms have also faced challenges with AI-generated inaccuracies in court. Meanwhile, investment in AI-driven legal automation continues to grow, exemplified by startups like Harvey seeking substantial funding despite these ongoing pitfalls.