In a significant development within the ongoing legal battle between Anthropic and several music publishers, including Universal Music Group, Anthropic’s lawyer acknowledged using an erroneous citation generated by the company’s AI chatbot, Claude. This concession was made during a lawsuit proceeding in Northern California, where that lawsuit has focused on the issue of copyright infringement.
The erroneous citation was the product of a phenomenon called “hallucination” by Claude, the AI created by Anthropic. The chatbot produced a citation that included “an inaccurate title and inaccurate authors,” leading to confusion and further complications in the case. Despite conducting a “manual citation check,” Anthropic’s legal team failed to detect this particular error, as well as other inaccuracies caused by Claude’s outputs.
Olivia Chen, an expert witness for Anthropic, has been widely attacked. The allegations include her previously using Claude to reference made-up articles during her testimony. California Judge Susan van Keulen swung the gavel. She ordered Anthropic to address the specific and egregious claims regarding their reliance on invalid citations in court filings.
“This is an honest citation mistake and not a fabrication of authority.” – Anthropic
The lawsuit against Anthropic represents only one of the more high-profile disputes between copyright owners and technology platforms. These battlegrounds are the result of the alleged unlawful use of creative material in the training of generative AI technologies. This continued trend poses significant implications for intellectual property rights with the continued, rapid incorporation of AI technologies into all aspects of industry.
Leave a Reply