Behind the doors of El Cajon’s roughly 30-member investigations unit, detectives have been quietly putting a new partner to work: an artificial intelligence tool that promises to chew through piles of calls, photos and text messages from a single case in a fraction of the usual time. For now, it is just a one-case trial, not a full-blown purchase, but the test has already drawn some sharp lines. Investigators like the idea of shrinking digital backlogs. Civil-liberties advocates see creeping surveillance and potential bias. County leaders are watching from the sidelines, clipboards ready.
According to The San Diego Union-Tribune, El Cajon is running software from startup Longeye on a single investigation and has stopped short of rolling it out across the force. The Sheriff’s Office, for its part, is in a wait-and-see mode as it evaluates the same product. City officials have framed the experiment as a limited test rather than a shopping trip, while detectives told the paper they hope quicker searches through digital evidence will free them up for more in-depth follow-up work. The Union-Tribune reports that agencies across the region are weighing whether promised efficiency gains are worth the legal and privacy trade-offs.
Longeye, a San Francisco-based startup, relies on large-language-model techniques that let investigators type plain-English questions about uploaded evidence such as audio, transcripts, images and files instead of hunting through folders by hand, the Washington Post reported. The Post profiled early adopters and found departments using the tool to speed up case reviews that once dragged on for weeks. Investors and rival vendors have taken notice as short-staffed agencies consider rewriting their investigative workflows around tools like this.
Company executives say the platform points directly back to the underlying recordings and documents instead of spitting out predictive scores. Critics counter that if you automate analysis on top of messy, biased records, and do it without strong oversight, you can still turbocharge the errors.
Regulation And Privacy Questions
California has moved faster than many states on rules for automated systems. The state privacy regulator finalized new CCPA guidance on automated decision-making technology last year that stresses transparency, risk assessments and keeping vendors on the hook. Legal alerts note that the rules will mean new documentation and governance requirements for systems that substantially replace human decision-making, and local officials have been paying attention.
County supervisors have directed staff to study AI policy and look at adding extra protections to large IT contracts, according to analysis from JDSupra. That emerging framework is shaping up as the backdrop for any move San Diego-area departments might make on tools like Longeye.
Why Some Detectives Are Tempted
For current and former law enforcement leaders who have tried the tool, the attraction is simple: hours saved on drudgery can be redirected to actual police work. Time that used to be spent listening to endless recordings or scrolling through transcripts can go instead toward interviews and fieldwork.
In an interview with the Livermore Vine, retired police chief Denton Carlson, now a public-safety ambassador for Longeye, described demos where the platform surfaced a crucial snippet of audio in minutes, a task that would once have taken detectives days. Examples like that help explain why cash-strapped departments are at least hearing out the sales pitch, even as privacy advocates press hard for strict limits.
What’s Next In San Diego
Local officials say the next phase will be about oversight as much as speed. El Cajon has not signed on for a full deployment after its pilot, and county law enforcement leaders say they are proceeding cautiously while vendors make their rounds. The San Diego Union-Tribune reports that Longeye is negotiating contracts around the country and pitching departments on rapid evidence review, while prosecutors and public defenders are starting to ask how AI-generated summaries will be checked in court. Officials say any procurement decisions will likely hinge on written guarantees about auditing, data handling and human review.
For now, the El Cajon pilot is a small but weighty test case. It sits at the intersection of shrinking detective ranks and rising legal and privacy expectations under state rules. City and county leaders say they plan to stack the pilot’s results against new state guidance and their own policy work before they sign off on anything broader.
