Anthropic's Lawyer Apologizes After Claude Hallucinates Legal Citation

In a recent development, a lawyer representing Anthropic acknowledged the use of an incorrect citation generated by the company’s Claude AI chatbot during its ongoing legal dispute with music publishers. The admission came via a court filing in a Northern California courtroom on Thursday.
Anthropic stated in the filing, first reported by Bloomberg, that Claude had fabricated the citation with both an incorrect title and author details. The company's lawyers admitted that their routine citation checks failed to detect the error, along with several other inaccuracies caused by Claude's "hallucinations."
Apologizing for the oversight, Anthropic described the incident as "an honest citation mistake and not a fabrication of authority."
Earlier this week, attorneys for Universal Music Group and other music publishers accused Anthropic’s expert witness, employee Olivia Chen, of referencing fabricated articles in her testimony, allegedly using Claude. Following this, Federal Judge Susan van Keulen requested a response from Anthropic regarding these claims.
The lawsuit brought forth by the music publishers is part of multiple conflicts between copyright holders and tech firms over the alleged misuse of their content to develop generative AI tools.
This marks yet another instance where the use of AI in legal proceedings has led to unintended consequences. Earlier this week, a California judge criticized two law firms for submitting "phony AI-generated research" in his court. In January, an Australian attorney was exposed for utilizing ChatGPT in drafting court documents, resulting in flawed citations.
Despite such incidents, startups continue to attract significant funding to automate legal processes. Harvey, which employs generative AI models to aid lawyers, is reportedly in discussions to secure over $250 million at a valuation exceeding $5 billion.
Techcrunch event ### Join us at TechCrunch Sessions: AI
Reserve your spot at our premier AI industry event featuring speakers from OpenAI, Anthropic, and Cohere. For a limited period, tickets are priced at just $292 for a full day of expert presentations, workshops, and valuable networking opportunities.
Exhibit at TechCrunch Sessions: AI
Book your展位 at TC Sessions: AI and showcase your innovations to 1,200+ decision-makers without breaking the bank. Available until May 9 or while spots last.
Berkeley, CA | June 5 REGISTER NOW
Related article
Anthropic Expands Compute Partnerships with Google and Broadrom
AI research lab Anthropic announced on Monday a new agreement with Google and Broadcom to significantly boost the processing and computational power behind its Claude AI models. This restructuring of its compute partnerships arrives as demand for its
Claude Gains Ground on ChatGPT as Users Migrate
Following a series of controversies involving ChatGPT and its parent company OpenAI, a growing number of users are migrating to Claude.The turning point occurred after Anthropic, Claude's creator, declined a Department of Defense request to utilize i
What Anthropic's Pentagon Standoff Means for National Security
The past two weeks have been dominated by a public standoff between Anthropic CEO Dario Amodei and Defense Secretary Pete Hegseth, centering on the military's application of AI technology.Anthropic has established policies prohibiting its AI models f
Related Special Topic Recommendations
Comments (2)
0/500
Kann jemand erklären, warum selbst Top-LLMs noch Halluzinationen produzieren? Besonders bei Rechtsthemen ist das echt kritisch. Die Entschuldigung ist ein Schritt, aber wie werden solche Fehler in Zukunft verhindert? 🤔 Bin gespannt, ob sich da bald branchenweite Standards durchsetzen.

In a recent development, a lawyer representing Anthropic acknowledged the use of an incorrect citation generated by the company’s Claude AI chatbot during its ongoing legal dispute with music publishers. The admission came via a court filing in a Northern California courtroom on Thursday.
Anthropic stated in the filing, first reported by Bloomberg, that Claude had fabricated the citation with both an incorrect title and author details. The company's lawyers admitted that their routine citation checks failed to detect the error, along with several other inaccuracies caused by Claude's "hallucinations."
Apologizing for the oversight, Anthropic described the incident as "an honest citation mistake and not a fabrication of authority."
Earlier this week, attorneys for Universal Music Group and other music publishers accused Anthropic’s expert witness, employee Olivia Chen, of referencing fabricated articles in her testimony, allegedly using Claude. Following this, Federal Judge Susan van Keulen requested a response from Anthropic regarding these claims.
The lawsuit brought forth by the music publishers is part of multiple conflicts between copyright holders and tech firms over the alleged misuse of their content to develop generative AI tools.
This marks yet another instance where the use of AI in legal proceedings has led to unintended consequences. Earlier this week, a California judge criticized two law firms for submitting "phony AI-generated research" in his court. In January, an Australian attorney was exposed for utilizing ChatGPT in drafting court documents, resulting in flawed citations.
Despite such incidents, startups continue to attract significant funding to automate legal processes. Harvey, which employs generative AI models to aid lawyers, is reportedly in discussions to secure over $250 million at a valuation exceeding $5 billion.
Techcrunch event ### Join us at TechCrunch Sessions: AI
Reserve your spot at our premier AI industry event featuring speakers from OpenAI, Anthropic, and Cohere. For a limited period, tickets are priced at just $292 for a full day of expert presentations, workshops, and valuable networking opportunities.
Exhibit at TechCrunch Sessions: AI
Book your展位 at TC Sessions: AI and showcase your innovations to 1,200+ decision-makers without breaking the bank. Available until May 9 or while spots last.
Berkeley, CA | June 5 REGISTER NOW
Anthropic Expands Compute Partnerships with Google and Broadrom
AI research lab Anthropic announced on Monday a new agreement with Google and Broadcom to significantly boost the processing and computational power behind its Claude AI models. This restructuring of its compute partnerships arrives as demand for its
Claude Gains Ground on ChatGPT as Users Migrate
Following a series of controversies involving ChatGPT and its parent company OpenAI, a growing number of users are migrating to Claude.The turning point occurred after Anthropic, Claude's creator, declined a Department of Defense request to utilize i
What Anthropic's Pentagon Standoff Means for National Security
The past two weeks have been dominated by a public standoff between Anthropic CEO Dario Amodei and Defense Secretary Pete Hegseth, centering on the military's application of AI technology.Anthropic has established policies prohibiting its AI models f
Kann jemand erklären, warum selbst Top-LLMs noch Halluzinationen produzieren? Besonders bei Rechtsthemen ist das echt kritisch. Die Entschuldigung ist ein Schritt, aber wie werden solche Fehler in Zukunft verhindert? 🤔 Bin gespannt, ob sich da bald branchenweite Standards durchsetzen.





Home






