No known arrests or prosecutions stemmed from San Diego's 7-year facial recognition program, which compiled 65K face scans; the program ended Dec. 31, 2019
DJ Pangburn / Fast Company : Tweets: @lizjosullivan , @donttag , @liz_in_shanghai , @matt_cagle , and @rogre Tweets: @lizjosullivan : People often ask me why I really want to ban facial recognition when it could be so useful for safety. What they don't realize is that there's absolutely no proof that it enhances safety from violent crime. There IS proof, however, that it fails a lot... https://www.fastcompany.com/ ... @donttag : It's banned in California now, but San Diego experimented with facial https://www.fastcompany.com/ ... “We will be terminating the usage of the [TACIDS-enabled] devices, the only biometric technology we will be utilizing is fingerprint scanners.” Liza Lin / @liz_in_shanghai : San Diego's massive, 7-year experiment with facial recognition technology appears to be a flop - since 2012, law enforcement collected as many as 65,000 face scans and tried to match it to a database. https://www.fastcompany.com/ ... Matt Cagle / @matt_cagle : California's new facial recognition law (AB 1215) notches its first victory: San Diego has terminated a massive and strikingly ineffective face scanning program. https://www.fastcompany.com/ ... More on AB 1215 & the campaign to pass it here: https://www.aclu.org/... Roberto Greco / @rogre : “San Diego's love affair with surveillance technology extends beyond just TACIDS... tattoo matching software... license plate reader system... drones to photograph crime scenes, locate homeless encampments, and assist with emergency response.” https://www.fastcompany.com/ ...
Context & Ripple Effects
San Diego spent seven years building a biometric database — roughly 65,000 face scans via TACIDS-enabled devices — and shut it down on Dec. 31, 2019 with no known arrests or prosecutions to show for it. The timing matters: California enacted AB 1215 around the same window, banning facial recognition, so the city ended its program just as the state foreclosed restarting one.
The zero-yield record lands in the middle of an active fight over efficacy claims. Activists arguing against renewals — like the campaign to end facial recognition in Detroit — now have a full-program case study where the promised safety benefit never materialized, while later reporting shows other departments kept using the tech anyway.
First-order effects
- San Diego police fall back to fingerprint scanners only, abandoning face-scan capability entirely rather than replacing the TACIDS devices.
- Vendors selling TACIDS-style biometric hardware lose a seven-year municipal deployment whose outcome — zero known arrests from 65,000 scans — undercuts the core sales pitch.
Second-order effects
- Cities still under contract, most visibly Detroit ahead of its renewal date, face activist campaigns armed with San Diego's no-results record as direct evidence that the technology does not deliver the violent-crime prevention vendors claim.
- Jurisdictions that keep the technology despite weak results — as Buenos Aires' subway system did through wrongful detentions, or New Orleans' secret network of 200+ cameras operating against a city ordinance — widen the gap between what councils approve and what police actually run.
Third-order effects
- If the pattern holds, the debate shifts from accuracy rates to demonstrated outcomes: bans like AB 1215 spread where programs can't show arrests, while departments that continue use do so outside approved channels, pushing oversight toward audits and enforcement of local ordinances rather than procurement decisions.
- A two-track landscape emerges — jurisdictions that ban after failed pilots versus departments that adopt regardless of evidence, echoing the Washington Post finding of arrests made on facial-recognition matches alone across 15 departments — leaving residents' exposure dependent on which side of that split their city falls on.
The trend: Municipal facial recognition is splitting into jurisdictions that terminate programs when efficacy data comes up empty and departments that keep scanning covertly, making enforcement of local controls the next battleground.