In a recent New York court case, an expert witness’s reliance on Microsoft’s Copilot chatbot to estimate damages in a real estate dispute has been met with criticism from the presiding judge. This is just the latest example of legal professionals and experts being embarrassed by relying on AI tools.
Ars Technica reports that New York Judge Jonathan Schopf, has called attention to the potential dangers of expert witnesses using AI tools, such as Microsoft’s Copilot chatbot, to provide testimony in court cases. The issue came to light during a real estate dispute involving a $485,000 rental property in the Bahamas, which had been included in a trust for a deceased man’s son.
The case revolved around the executrix and trustee, the deceased man’s sister, who was accused of breaching her fiduciary duties by delaying the sale of the property while using it for personal vacations. To prove damages, the surviving son’s legal team relied on expert witness Charles Ranson to calculate the difference between the potential sale price in 2008 and the actual sale price in 2022.
However, Ranson, who specialized in trust and estate litigation, lacked relevant real estate expertise. To compensate for this, he turned to Copilot AI to assist in his calculations. During his testimony, Ranson was unable to recall the specific prompts he used to arrive at his damages estimate or cite any sources for the information he obtained from the chatbot. He also admitted to having a limited understanding of how Copilot functions and generates outputs. Despite this, Ranson adamantly defended the use of Copilot and other AI tools for drafting expert reports, claiming it to be a generally accepted practice in the field of fiduciary services.
In response to Ranson’s testimony, Judge Schopf took the time to experiment with Copilot himself, attempting to generate the same estimates that Ranson had provided. However, he found that the chatbot produced slightly different answers each time, even when given the same query. This inconsistency raised concerns about the reliability and accuracy of Copilot-generated evidence in court proceedings.
Copilot itself, when questioned by the court about its accuracy and reliability, responded that its outputs should always be verified by experts and accompanied by professional evaluations before being used in court. This sentiment was echoed by Judge Schopf, who noted that the developers of Copilot recognize the need for human supervision to verify the accuracy of both the input information and the generated output.
As a result of this case, Judge Schopf has called for lawyers to disclose the use of AI in their cases to prevent chatbot-generated inadmissible testimony from disrupting the legal system. He emphasized that while AI is becoming increasingly prevalent across many industries, its mere presence does not automatically make its results admissible in court.
Ultimately, Judge Schopf found no breach of fiduciary duty in the case, rendering Ranson’s Copilot-assisted testimony on damages unnecessary. The judge denied all of the son’s objections and future claims, noting that Ranson’s testimony showed a lack of thorough analysis, the use of an incorrect time period for damages, and a failure to consider obvious elements in his calculations.
Read more at Ars Technica here.
Lucas Nolan is a reporter for Breitbart News covering issues of free speech and online censorship.
Read the full article here