Source: National Law Review

The U.S. Advisory Committee on Federal Rules of Evidence has proposed amendments to address the use of AI-generated evidence in litigation. Key changes include expanding Rule 901 to require detailed authentication of AI-generated outputs and introducing Rule 707, which applies the same standards used for expert testimony to AI outputs.

Under the proposed rules, proponents of AI-generated evidence must provide details about the training data, software, and reliability of the AI system. Additionally, any evidence suspected of being altered by AI will only be admissible if it is "more likely than not authentic."

The amendments aim to address risks such as analytical errors, bias, and lack of interpretability in AI systems. By setting stricter standards for the admissibility of AI outputs, the rules emphasize the importance of transparency and reliability in courtroom evidence, ensuring that technological advancements align with legal principles.