Police departments are increasingly adopting generative AI tools for report writing, exemplified by the introduction of Draft One by Axon earlier this year.

Oklahoma City's police department is among the few experimenting with AI-generated reports, generating mixed reactions from officers and legal experts.

Sgt. Matt Gilmore from Oklahoma City praised the AI-generated report for its accuracy and flow, noting it included details he had overlooked.

The introduction of Draft One aims to significantly reduce report-writing time by 30 to 45 minutes per report, allowing officers to focus more on community engagement.

Axon's CEO Rick Smith highlighted the positive reception of Draft One among officers, who prefer focusing on policing rather than tedious paperwork.

As police departments adopt technologies like AI, they are expected to adapt their communication styles to improve the accuracy of AI outputs.

Concerns about the reliability of generative AI in high-stakes situations have been raised, particularly regarding potential biases and issues of data privacy.

Legal scholar Andrew Ferguson emphasizes the need for public discussion on the implications of AI in law enforcement, particularly regarding the potential for false information in reports.

Experts warn that increased reliance on AI tools may lead to less meticulous report writing, as officers may become overly dependent on automated processes.

Currently, Draft One is limited to minor incident reports that do not involve arrests, as advised by local prosecutors.

Experts caution that AI applications in law may generate inaccuracies, with studies indicating significant error rates in AI-generated legal information.

Draft One utilizes Microsoft's Azure OpenAI platform to transcribe audio from police body cameras and generate draft narratives based strictly on those transcripts.