AI Frontier Institute

Sets rules and procedures for the admissibility of evidence created or processed by artificial intelligence.

View primary source →

New York Assembly Bill 1338 (2025) amends the Criminal Procedure Law and the Civil Practice Law and Rules to establish admissibility standards for evidence created or processed by artificial intelligence in both criminal and civil proceedings. Under new Criminal Procedure Law § 60.80, AI-created evidence may not be admitted in criminal cases unless it is substantially supported by independent, admissible evidence and the proponent proves the reliability and accuracy of the specific AI use. AI-processed evidence requires only the reliability and accuracy showing, without the independent-evidence requirement. New Civil Practice Law and Rules § 4552 mirrors these standards for civil proceedings. The bill defines 'created' evidence as AI producing new information not present in or reasonably deducible from existing information, and 'processed' evidence as AI drawing conclusions from analysis or transformation of existing information where those conclusions are not reasonably deducible. To establish reliability and accuracy, the proponent must present a qualified expert who testifies that: the AI use has been validated through rigorous scientific or technical testing; the AI has been tested in environments similar to the proceeding context; and the AI has not been exposed to variables with a substantial probability of causing materially inaccurate results. Courts may impose protective measures when expert testimony would reveal trade secrets, privileged government information, or sensitive law enforcement techniques. The act takes effect January 1 of the year following enactment.

← Back to the tracker