In a recent filing, an attorney in Texas allowed ChatGPT, an AI language model, to supplement his legal research but all of the six cases and relevant precedent provided were entirely fabricated. In response to this, Judge Brantley Starr has included a requirement that any attorney appearing in his court must declare and verify that no portion of the filing was drafted by generative artificial intelligence, or if it was, that it was verified “by a human being”. At the federal site for Texas’s Northern District, the judge has the opportunity to set specific rules for his courtroom, hence the “Mandatory Certification Regarding Generative Artificial Intelligence” was added. The certification requires attorneys appearing in court to file a certificate declaring that any language drafted by generative artificial intelligence was checked by a human being for accuracy. The form for lawyers to sign also specifies that quotations, citations, paraphrased assertions, and legal analysis are covered by this declaration. Although AI could potentially assist with summarising and finding precedent, it is subject to hallucinations and bias, hence the judge’s ruling.
Steven Schwartz is the attorney who allowed ChatGPT to supplement his legal research in a recent federal filing and has since expressed regret. It is unclear whether the inclusion of the Mandatory Certification Regarding Generative Artificial Intelligence is in response to this specific case.
ChatGPT is an AI language model designed by OpenAI that is able to understand and generate human-like text. It was designed to complete specific prompts such as answer questions or continue text, making it useful for purposes such as chatbots, content generation, and more.
