The general counsel for the Oregon State Bar said fabricated cases and citations have become more common among lawyers and people representing themselves.
Penalties for AI Fabrications Two Oregon lawyers were recently fined a total of $110,000 for submitting legal documents containing fabricated cases and citations generated by Artificial Intelligence, indicating a growing problem. Prevalence and Definition of AI Hallucinations According to Ankur Doshi, general counsel of the Oregon State Bar, there have been around five identified court filings in Oregon and approximately 900 nationally that contained AI 'hallucinations.' These occur when generative AI programs like Claude or ChatGPT produce inaccurate, misleading, or entirely invented information. AI Misuse by Lawyers and Self-Represented Individuals Doshi highlighted that AI misuse in legal cases is not limited to lawyers but also includes non-lawyers representing themselves (pro se) who use these tools to draft their own pleadings, potentially facing court sanctions such as fines. Disciplinary Actions and Transparency Requirements The discipline for lawyers misusing AI varies based on 'aggravating' or 'mitigating' factors, particularly the lawyer's transparency. Disclosure of AI use might result in a minor reprimand, while concealment could lead to severe penalties, including fines, suspension, or disbarment. The Oregon State Bar requires lawyers to be competent in any tools they use, including understanding AI's propensity to generate false information. Balancing Efficiency with Human Oversight Many lawyers are willing to use AI due to its potential for substantial time savings and efficiencies in reviewing and drafting legal work. However, Doshi emphasizes that lawyers must thoroughly understand the AI tools they employ and always apply human oversight to verify the accuracy of AI-generated content. Threat to the Legal System's Integrity AI mistakes, such as fabricated cases and statements of law, can lead to increased costs for courts due to the extra work required to review false information. More critically, Doshi warns that such errors directly undermine the integrity of the precedent-based legal system.
Penalties for AI Fabrications
Two Oregon lawyers were recently fined a total of $110,000 for submitting legal documents containing fabricated cases and citations generated by Artificial Intelligence, indicating a growing problem.
Prevalence and Definition of AI Hallucinations
According to Ankur Doshi, general counsel of the Oregon State Bar, there have been around five identified court filings in Oregon and approximately 900 nationally that contained AI 'hallucinations.' These occur when generative AI programs like Claude or ChatGPT produce inaccurate, misleading, or entirely invented information.
AI Misuse by Lawyers and Self-Represented Individuals
Doshi highlighted that AI misuse in legal cases is not limited to lawyers but also includes non-lawyers representing themselves (pro se) who use these tools to draft their own pleadings, potentially facing court sanctions such as fines.
Disciplinary Actions and Transparency Requirements
The discipline for lawyers misusing AI varies based on 'aggravating' or 'mitigating' factors, particularly the lawyer's transparency. Disclosure of AI use might result in a minor reprimand, while concealment could lead to severe penalties, including fines, suspension, or disbarment. The Oregon State Bar requires lawyers to be competent in any tools they use, including understanding AI's propensity to generate false information.
Balancing Efficiency with Human Oversight
Many lawyers are willing to use AI due to its potential for substantial time savings and efficiencies in reviewing and drafting legal work. However, Doshi emphasizes that lawyers must thoroughly understand the AI tools they employ and always apply human oversight to verify the accuracy of AI-generated content.
Threat to the Legal System's Integrity
AI mistakes, such as fabricated cases and statements of law, can lead to increased costs for courts due to the extra work required to review false information. More critically, Doshi warns that such errors directly undermine the integrity of the precedent-based legal system.
