POL Biannual Surveillance Report 2026

1. Please (a) describe the technology or technologies and (b) make a clear notation of which policy each technology corresponds to.

Automated License Plate Readers (ALPR), approved by Board of Supervisors on December 12, 2023:
ALPR cameras capture images of vehicles and their license plates to search against hotlists of stolen or wanted vehicles. The Department operates 400 fixed Flock Safety ALPR cameras, first installed in March 2024. Members may also run plate searches through Thomson Reuters CLEAR, whose license plate data comes from the Motorola/Vigilant network. Motorola vehicle-mounted ALPR units were installed on two Department vehicles for testing beginning June 2026.

ShotSpotter, approved by the Board of Supervisors August 4, 2021:
ShotSpotter (SoundThinking, Inc.) is an acoustic gunshot detection system. Sensors mounted on rooftops and light structures detect loud impulsive sounds and triangulate their location. A SoundThinking analyst reviews each event, and confirmed gunfire is published to officers' phones and to the Department of Emergency Management's 911 dispatch center with a location, time and round count. Coverage areas fall within the Southern, Bayview, Mission, Northern, Ingleside and Tenderloin police districts.

Non-City Entity Cameras, approved by Board of Supervisors on November 6, 2022:
The policy governs the Department's access to security cameras owned by private businesses, residents and other nonCity entities. The Department does not own or operate these cameras. Access to data is by consent of the owner, or by subpoena, search warrant or court order. Historical footage may be requested for criminal investigations. Temporary live monitoring is limited to 24 hours, requires captain-level approval and a signed Department consent form (Form 619) from the camera owner, and may not be recorded. 

2. How was the surveillance technology or technologies used by your department during the reporting period? Provide a 3-4 sentence description for each technology listed.

ALPR: Members used ALPR to identify and locate vehicles connected to criminal investigations, including stolen vehicles, vehicles linked to robberies, burglaries and shootings, and missing persons. As required by Department Notice 26-011, members must enter an incident number, dispatch call number or stated reason for every search, and hotlist alerts notify members when a stolen or wanted vehicle passes a camera. Under state law, ALPR data may only be shared only with California public agencies.

ShotSpotter: ShotSpotter alerted patrol officers and the Department of Emergency Management's 911 dispatch center to gunfire, often where no one called 911. Officers responded to the alert locations. Activation data, including time, location, round count and a short audio clip, was provided to investigators, to the District Attorney's Office for prosecution, and to Crime Gun Investigations Center partners for official law enforcement purposes.

Non-City Entity Cameras: Members requested historical footage from businesses and residences as part of criminal investigations. From January 2025 through June 2026, the Department made 21 requests for temporary live monitoring, all 21 were approved by the camera owners, and the resulting 21 operations were concentrated in the Tenderloin, Central and Southern districts. Live monitoring was used mainly for felony and misdemeanor investigations. It was also used twice in response to community complaints of criminal activity and once during a significant event, which Department policy defines as a large or high-profile event in the City requiring enhanced situational awareness and coordination of police resources.

3. Has the surveillance technology been effective at achieving its identified purpose?

Yes.

4. In 4-10 sentences, please explain how the technology has or has not been effective.

ALPR: ALPR has become one of the Department's most-used investigative tools. It allows investigators to locate stolen vehicles and vehicles connected to violent and property crimes within minutes rather than days. Reported vehicle crime has declined sharply since the network was installed in 2024. The Department's crime statistics are drawn from police incident reports, which include crimes reported by the public by phone, online or in person, as well as crimes officers discover on patrol. Based on those reports, motor vehicle thefts fell 44 percent, burglaries fell 29 percent and car breakins fell 43 percent in 2025 compared with 2024. Through September 6, 2026, reported motor vehicle thefts were down a further 22 percent and burglaries a further 23 percent compared with the same period in 2025. Multiple factors contributed to these declines, and ALPR has been a key investigative tool in many of those cases. 

ShotSpotter: ShotSpotter makes the Department aware of gunfire it would not otherwise know about. A 2019 analysis by the Department of Emergency Management found that only about 15 percent of gunfire incidents in San Francisco produced a 911 call. Alerts give officers a precise location within about 60 seconds, so they can reach victims faster, protect the scene and recover evidence. Activation data is also shared with Crime Gun Investigations Center partners to support firearm investigations. Gun violence continued to fall during the reporting period. The Department's crime statistics are drawn from police incident reports, which include crimes reported by the public and crimes officers discover on patrol. Based on those reports, shootings fell 16 percent and homicides fell 20 percent in 2025 compared with 2024. The Department's 2025 year-end crime report found that 28 homicides was the lowest annual total in San Francisco since 1954. The decline has continued in 2026: through September 6, the number of reported gun violence victims was down 19 percent from the same period in 2025. 

Non-City Entity Cameras: Access to privately owned cameras allows the Department to investigate crimes without owning or installing cameras of its own. Live monitoring has been used narrowly and every request was voluntarily approved by the camera owner. From January 2025 through June 2026, live monitoring operations led to arrests in 18 of 21 operations. Arrests included homicide, aggravated assault, robbery, possession of a firearm by a prohibited person, narcotics sales, burglary and auto theft. Historical footage from private cameras is also routinely used to identify suspects and corroborate witness accounts.

4.a. Provide quantitative data to support your response. This should include crime statistics for the radius where the
technology operates if that was a motivating factor in acquiring the surveillance technology. 

ALPR: The Department operates 400 fixed Flock Safety ALPR cameras and two vehicle-mounted Motorola ALPR cameras. The Department's crime statistics are drawn from police incident reports, which include crimes reported by the public and crimes officers discover on patrol. From January 1 through November 16, 2025, the Department recorded 2,730 motor vehicle thefts, compared with 4,907 in the same period of 2024, a 44 percent decrease, and 3,300 burglaries, compared with 4,611, a 28 percent decrease. For the full year, the Department's 2025 year-end crime report found that car break-ins fell 43 percent compared with 2024. From January 1 through September 6, 2026, the Department recorded 1,721 motor vehicle thefts, compared with 2,220 in the same period of 2025, a 22 percent decrease. Over the same period, burglaries fell from 2,682 to 2,073 (23 percent), larceny thefts fell from 11,915 to 9,729 (18 percent), robberies fell from 1,152 to 946 (18 percent), and total violent crimes fell from 2,800 to 2,554 (9 percent). The 2026 figures are preliminary and subject to revision. 

ShotSpotter: The Department's crime statistics are drawn from police incident reports, which include crimes reported by the public and crimes officers discover on patrol. According to the Department's 2025 year-end crime report, shootings fell 16 percent in 2025 compared with 2024. Homicides fell from 35 in 2024 to 28 in 2025, a 20 percent decrease and the lowest annual total since 1954. From January 1 through September 6, 2026, there were 65 gun violence victims, compared with 80 in the same period of 2025, a 19 percent decrease, and 49 non-fatal shooting incidents, compared with 57, a 14 percent decrease. The 2026 figures are preliminary and subject to revision.

Non-City Entity Cameras: From January through June 2026, the Department made 11 requests for live monitoring, the camera owners approved all 11, and 10 of the 11 operations led to an arrest; officers spent a total of 40 hours and 58 minutes monitoring, and 37 members viewed the footage. From July through December 2025, the Department made 7 requests, the camera owners approved all 7, and 5 operations led to an arrest; officers spent 14 hours and 57 minutes monitoring, and 10 members viewed the footage. From January through June 2025, the Department made 3 requests, the camera owners approved all 3, and all 3 operations led to an arrest; officers spent 21 hours monitoring, and 4 members viewed the footage. In total, from January 2025 through June 2026, the Department made 21 live monitoring requests, all 21 were approved, and 18 operations led to an arrest, with 76 hours and 55 minutes of total monitoring. Census-tract crime statistics for the month before and after each operation are reported in the Department's biannual live monitoring reports to the Police Commission and the Board of Supervisors.

5. How many complaints or concerns has your department received from the public about the surveillance technology?

ALPR: The Department received three concerns from the public about ALPR during the reporting period, which are summarized in Question 6.
ShotSpotter: The Department did not receive any complaints or concerns from the public about ShotSpotter during the reporting period.
Non-City Entity Cameras: The Department did not receive any complaints or concerns from the public about its access to non-City entity cameras during the reporting period.

6. Please summarize the complaints or concerns which your department received about the surveillance technology. 

ALPR:

• In September 2025, the Electronic Frontier Foundation and the American Civil Liberties Union sent the Department a demand letter and public records request urging a full audit and new compliance protocols.
• In November 2025, a Request for Inquiry was filed with the Human Rights Commission alleging a violation of the City's
  Sanctuary Ordinance (Administrative Code Chapter 12H). The Human Rights Commission's September 2026 report
  found no violation but made five recommendations, which the Department responded to in writing on August 26, 2026.
• In December 2025, a resident filed a federal lawsuit alleging that the City's use of Flock cameras violates the Fourth
  Amendment. The lawsuit was dismissed.

ShotSpotter: No complaints or concerns were received.

Non-City Entity Cameras: No complaints or concerns were received.

7. How many complaints or concerns has your department received from the public about the receipt of information from face recognition technology?
ALPR: None.
ShotSpotter: None.
Non-City Entity Cameras: None.

Log of inadvertent receipt of face recognition information (Administrative Code Section 19B.2(e)): No complaints were received, but the Department logged the following receipt. On August 7, 2025, an investigator in the Department's Strategic Investigations Unit sent a routine bulletin with an image of a shooting suspect to Bay Area law enforcement agencies, as Department policy requires. An analyst with the Northern California Regional Intelligence Center responded with a possible identification obtained through face recognition. The Department did not request or solicit this information, and no Department member used face recognition technology. The investigator independently corroborated the suspect's identity through other evidence and disclosed the source of the Regional Intelligence Center's information to the court in a search warrant affidavit. The receipt was documented in a memorandum to the Commanding Officer of Strategic Investigations dated September 4, 2025. This receipt is not associated with the ALPR, ShotSpotter, or Non-City Entity Camera technologies covered in this report.

8. Please summarize the complaints or concerns which your department received about the receipt of information from face recognition technology.
ALPR: Not applicable.
ShotSpotter: Not applicable.
Non-City Entity Cameras: Not applicable.
 

9. Were there any violations of the Surveillance Technology Policy or Surveillance Impact Report, reported through community members, non-privileged internal audits, or through other means in the last year?
Yes

9.a. How many violations have there been over the last year?
3 

9.b. Please describe each violation and document what the department did to respond to violation – in terms of correction, public disclosure, and discipline of involved parties.
1. An investigation by the San Francisco Police Department's own Internal Affairs Division found that an officer searched the Flock system for a spouse's stolen vehicle and posted Flock images on social media, in violation of the Department's conflict-of-interest policy. The case was reported to the Police Commission in the Disciplinary Review Board's report for the third quarter of 2025.
2. During a routine compliance audit in May 2026, the Department found that the Northern California Regional Intelligence Center shared SFPD ALPR data access with Western States Information Network, whose analysts ran 299 queries of the Department's Flock network between May 1, 2025 and May 12, 2026 on behalf of federal and out-of-state agencies. These searches were not run by Department members, and none referenced immigration enforcement or reproductive health care. Chief Lew reported the matter to the Police Commission on June 17, 2026. The Department of Police Accountability and the Human Rights Commission were also notified.
3. Between approximately August 2024 and February 2025, a Flock feature called Nationwide Lookup allowed out-ofstate agencies to query the Department's network. Flock enabled the feature without the Department's knowledge or authorization; the Department never approved data sharing with any out-of-state or federal agency. Flock has since permanently disabled Nationwide Lookup for all California agencies. The Human Rights Commission's September 2026 report found no violation of the Sanctuary Ordinance by the Department.

10. Is your department requesting to modify the Surveillance Technology Policy or Policies covered in this report?
No.