Facing the Algorithm: AI in Police Reports and Its Impact on the Criminal Justice System
Speaker: Robert Thies, Investigator, Federal Public Defender – Eastern District of Missouri
Hosted by AILA and LJST | November 11, 2025
By: Caitlin Kennedy Downey, AILA Staff
AILA and LJST welcomed Robert Thies, Investigator with the Federal Public Defender’s Office for the Eastern District of Missouri, for a thought-provoking virtual talk on the growing role of generative AI in criminal justice.
Thies explored how police departments are beginning to use AI tools like Axon’s Draft One to generate first drafts of police reports directly from body camera transcripts. This introduces a new layer of complexity to questions of authorship, accuracy, and accountability. In many jurisdictions, these drafts are deleted once edited by officers, leaving no record of the AI’s contribution to the narrative.
The conversation highlighted critical risks associated with AI-generated evidence, including bias, transparency, and the “black box” nature of many AI systems, especially since these models are often trained on biased data. Thies discussed California’s emerging requirements that all AI-generated materials, including drafts, be retained and disclosed when used in official reporting.
Looking ahead, Thies introduced Rule 707, a proposed addition to the Federal Rules of Evidence that aims to establish a clear framework for validating AI-generated materials before they are introduced in court. His proposed standards emphasize validation, operation, and transparency to ensure that AI systems can be trusted, that their outputs are explainable, and that they can be examined by all parties.
“The first draft should always be yours,” Thies emphasized. “Otherwise, you’ve lost control of the narrative.”
The discussion, moderated by Department Chair of LJST, Adam Sitze, raised broader questions about trust, expertise, and the future of evidence—from predictive policing to the challenge of deepfakes and the erosion of human memory in the face of machine-generated narratives.
The event left participants reflecting on how essential human judgment remains in ensuring fairness, transparency, and truth.