Michael Cohen used Google Bard AI to generate fake legal case citations that his lawyer included in court filings, leading to judicial sanctions and highlighting the risks of AI hallucinations in legal practice.
Michael Cohen, Donald Trump's former lawyer and fixer, used Google Bard, an AI chatbot developed by Google, to conduct legal research for a motion seeking early termination of his supervised release following his 2018 conviction for campaign finance violations. Cohen mistakenly believed Bard was a 'supercharged search engine' rather than a generative AI service and provided his attorney David Schwartz with three fictitious legal case citations generated by the AI. Schwartz included these fake citations in his November 2023 court filing without verifying their authenticity. U.S. District Judge Jesse Furman discovered that none of the cited cases existed and ordered Schwartz to explain the inclusion of non-existent cases or face sanctions. Cohen admitted in December 2023 court papers that he had not kept up with emerging legal technology trends and did not realize that Google Bard could generate citations that looked real but were actually fabricated. This incident follows a similar case earlier in 2023 where lawyers Steven Schwartz and Peter LoDuca were fined $5,000 for using ChatGPT-generated fake citations in an airline injury lawsuit. The episode raises concerns about Cohen's credibility as he is expected to be a star witness in Manhattan District Attorney's criminal case against Trump regarding hush money payments.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed