Operation LASER & PredPol — Predictive Policing
The LAPD ran two separate predictive systems: Operation LASER, which scored individuals as chronic offenders from 2011, and PredPol, which forecast property-crime hotspots. Its own Inspector General found in March 2019 that scoring was applied inconsistently and the crime effect could not be measured. The scoring tool ended in April 2019; PredPol was cut in April 2020 on budget grounds.
Completed
📅 2011–2020
🤖 Uses AI
✔ Editor-reviewed
📊 Impact
A randomised trial across the Foothill, North Hollywood and Southwest divisions found patrols directed by the ETAS algorithm behind PredPol cut crime volume by an average 7.4% against analyst-drawn hotspots (Mohler et al., JASA 2015), rated Promising by NIJ CrimeSolutions — though an LAPD officer and PredPol's co-founder were among the authors. In live operation the LAPD's own Inspector General reported it could not measure either programme's effect on crime.
🎓 Lesson
A department that cannot measure a tool cannot defend it. Los Angeles ran crime forecasting for nine years and its own watchdog still could not say what it had changed.
Evidence🔬 Peer-reviewed
Scale🏙️ City-wide
Sources
↗ LAPD Office of the Inspector General — Review of Selected Los Angeles Police Department Data-Driven Policing Strategies (8 Mar 2019, BPC #19-0072)The load-bearing document, and the reason the two programmes sit in one entry: the department's own watchdog reviewed Operation LASER (Chronic Offender component and LASER zones) and PredPol together, and reported to the Police Commission on 12 Mar 2019. Caution for a future check: the file is a 1.4 MB scan that did not yield machine-readable text in this review, so its findings are cited here through contemporaneous reports of it rather than quoted from the original.
↗ Government Technology / Los Angeles Times — LAPD to Halt Some Data Programs After Critical Audit (8 Apr 2019)Reporting of the Inspector General's findings: the chronic-offender labelling 'lacked oversight' and used 'inconsistent criteria'; 44% of designated chronic offenders had zero or one violent-crime arrest and about half had none for a gun-related crime; the effectiveness of the place-based component 'couldn't be determined'. Records that Moore had already suspended the tool in August 2018.
↗ Patch — LAPD to scrap controversial program to ID likely criminals (9 Apr 2019)Announcement date and Chief Michel Moore's own words: the tool 'did not bear the types of results that in the face of its criticism and the face of concerns relative to whether or not the underlying algorithm truly was fair and unbiased for us to continue going forward'. Also the two facts that keep this entry honest in both directions: the place-based LASER zones were kept and 'retooled' toward micro hot spots of about 500 by 500 feet rather than abolished, and the Inspector General found the racial makeup of those labelled chronic offenders 'comparable to the demographics of those arrested for violent crimes in the city'.
↗ BuzzFeed News — LAPD announces it is dropping PredPol because of the coronavirus pandemic (Apr 2020)Carries the LAPD memo of 15 April 2020 verbatim: 'The city's financial crisis, coupled with the impact of the COVID-19 pandemic, has resulted in the immediate freeze of new contractual agreements... Therefore, the Department will immediately discontinue the use of PredPol and its associated reports.' Also dates the LAPD's PredPol contract to 2011, and notes that campaigners dispute the budget explanation and credit their own pressure.
↗ MuckRock — Eight years in, LAPD can't measure PredPol's effect on crime (12 Mar 2019)Quotes the Inspector General directly — '[T]he OIG cautions against drawing strong conclusions from the available statistics' — and explains why: the recorded 'dosage', the time officers actually spent in a targeted area, was too imprecise to attribute any crime change to the programme.
↗ US Bureau of Justice Assistance — The Los Angeles Smart Policing Initiative: Reducing Gun-Related Violence through Operation LASEROrigin and structure, from the federal grant programme that paid for it: LASER was established in September 2011 in Newton Division with two distinct components, a chronic-offender component and a chronic-location component. This is the source that keeps the person-based and place-based halves separate.
↗ Mohler et al. — Randomized Controlled Field Trials of Predictive Policing, JASA 110(512), 2015The randomised trial behind the headline figure, run in three LAPD divisions (and separately in Kent, UK). Read with its conflicts visible: the author list includes an LAPD officer and P. Jeffrey Brantingham, PredPol's co-founder, so it is peer-reviewed but not arm's-length.
↗ NIJ CrimeSolutions — Program Profile: Predictive Policing Model in Los Angeles, Calif.Independent US Department of Justice assessment of that trial: rated 'Promising | One study', naming the Foothill, North Hollywood and Southwest divisions and the average 7.4% reduction in crime volume. Also the currency check — the page states the LAPD no longer uses the programme, and that the software later traded as PredPol and then Geolitica.
↗ Uchida & Swatt — Operation LASER and the Effectiveness of Hotspot Patrol: A Panel Analysis, Police Quarterly 16(3), 2013The strongest case the supporters had: a significant reduction in gun crime in Newton reporting districts against other divisions, found only where the chronic-offender and chronic-location components were applied together. Not independent — Uchida's firm was the programme's own Smart Policing Initiative research partner.
↗ Brayne — Big Data Surveillance: The Case of Policing, American Sociological Review 82(5), 2017Independent peer-reviewed ethnography of this specific department, from more than two and a half years of fieldwork and 75 interviews: discretionary risk judgements become numeric scores, data are used predictively rather than reactively, automated alerts widen the surveilled population, and people with no direct police contact enter the databases.
↗ Stop LAPD Spying Coalition — Data-Driven Policing (campaign archive)Advocacy source, labelled as such. The coalition's public-records campaign and its reports Before the Bullet Hits the Body (2018) and Automating Banishment (2021) forced much of the documentation into public view; it also argues the LAPD's successor framework, announced as the programmes wound down, is a rebrand rather than an end. That characterisation is the coalition's, not a verified finding.
📎 Cite this project
1001 Smart Cities (2026). “Operation LASER & PredPol — Predictive Policing” — Los Angeles, United States. The Smart City Atlas. https://1001smartcities.org/projects/la-predictive-policing/ (last verified 2026-07-26). Data: CC BY 4.0.
The underlying data is free to reuse with attribution — see the open data page.
🔗 Related projects
🇸🇬 TraceTogether & SafeEntry Digital Contact Tracing · Singapore
Two systems switched off on purpose: Singapore deleted its tracing data in 2024; LA dropped scoring in 2019 after its own watchdog audit.
🇮🇳 Citywide CCTV Rollout and Delhi Police Facial Recognition · New Delhi
Delhi's police told a court their face matching ran at 1-2% and kept going; LA measured nothing, and dropped the tool a month later.
🇭🇰 Multi-Functional Smart Lamppost Blueprint · Hong Kong
Two surveillance builds cut short: Hong Kong's lampposts were sawn down in 2019, LA's crime forecasting was cancelled by memo in 2020.
🇺🇸 Local Law 144: Bias Audits for Hiring Algorithms · New York City
Who audits the algorithm: New York makes employers hire an outside auditor; LAPD's was reviewed by the department's own Inspector General.
→ Open this project in the interactive atlas · → 📡 Data & IoT category · 🚩 Report an error