San Diego police explore AI-written reports as union chief challenges new spending

San Diego police want to explore whether AI-written reports could get officers back to service calls faster. But their union president is challenging the spending before any purchase. In an October 12 Voice of San Diego interview, Jared Wilson argued that the city should maintain its existing technology before buying more.

A funding objection, not a blanket rejection

Wilson, president of the San Diego police union, is not categorically against AI-assisted reporting. He told reporter Jakob McWhinney that careful implementation could help officers do more, particularly alongside strong community policing. His immediate objection is financial: he says San Diego lacks sustainable funding for technology it already uses.

He also questioned whether adopting another system would create a dependence the city could not easily reverse. Once officers rely on software, he warned, removing it could threaten response times and police capacity. That could leave the department stuck paying for technology even when its budget is strained.

His concern extends beyond software bills. Wilson said more policing technology can crowd out the neighborhood relationships he considers most effective. He favors giving officers time to become familiar figures in the communities they patrol, rather than having residents encounter an unfamiliar officer only after a crime.

Until the City of San Diego can maintain what it already has, we have to stop getting new stuff.

Jared Wilson, San Diego police union president, speaking to Voice of San Diego

An inquiry worth $500,000 to $3 million—not an award

The department issued its request for information on July 31, seeking software to help officers and investigators write police reports. Voice of San Diego’s October 9 report described an estimated potential contract value of $500,000 to $3 million. The document is an information-gathering and planning exercise, not a request for proposals.

Department spokesperson Ashley Nicholes says the work remains in an exploration phase, with no acquisition timeline. She described staffing shortages among front-line officers as a reason to look for ways to reduce time spent on tasks. The department’s stated goal is faster responses to service calls—not simply faster production of paperwork.

Control and consultation

• The request says officers must retain control over the content and structure of their reports.
• Nicholes said the department would reach out to elected officials and community members if it chooses to pursue the software.

The inquiry does not necessarily overturn the department’s earlier restriction. Its December policy prohibited officers from using non-department-approved generative AI programs to author reports. That distinction leaves room for an approved system, rather than establishing a permanent ban on every form of AI-assisted report writing.

A cleaner narrative still has to be accurate

California’s Senate Bill 524 required police departments to develop policies governing AI use, disclose when AI helped write reports, and retain body-camera footage and audio for future audits. Those requirements make disclosure and preservation of source material part of the oversight framework for AI-generated police paperwork.

Accuracy is a separate challenge from polished writing. Voice of San Diego cited a study using an Axon platform in which AI-generated police reports were rated significantly worse for accuracy. Study leader Ian Adams, a criminology professor and former officer, emphasized that prosecutors, defense attorneys, judges and juries rely on these documents to reconstruct events.

Khalid Alexander, president of Pillars of the Community, raised concerns about security and over-reliance. He questioned the assumption that AI would be less biased or more correct than people. His organization supports San Diegans affected by law enforcement, and he sees the proposed technology as taking humanity out of policing.

George Washington University law professor Andrew Ferguson offered a more qualified skepticism. Human-written reports also contain errors, he noted. But he warned that algorithmically changing a narrative could distort the facts used in criminal trials. His challenge to officials was to ask hard questions before deployment, rather than adopt technology and respond to problems afterward.