LAPD would delete nearly 12 million body camera videos under proposed policy change
-
Click here to listen to this article - Share via
See more from the L.A. Times in Google Search. Set us as preferred
- LAPD officials are proposing changes to the department’s data retention policy, seeking to purge millions of old records.
- The changes would require videos of police shootings and footage from investigations to be preserved.
- Police commissioners tabled a vote on the matter until the department addressed concerns.
The Los Angeles Police Department is seeking a policy change that would allow millions of videos collected from officers’ body-worn and dashboard-mounted cameras to be deleted, leaving oversight officials worried that useful footage might be lost in the purge.
In a presentation to the Board of Police Commissioners on Tuesday, the LAPD’s chief information officer, John Furay, detailed new data retention guidelines that would allow certain footage to be destroyed after five years.
Exceptions would be made for videos from all police shootings, as well as any potential evidence in a court case or internal investigation.
Under the current policy, all footage collected by the department is retained indefinitely.
In a video clip that has circulated widely online, Deputy Chief Alan Hamilton is heard making allegations against activist Jason Reedy during a Martin Luther King Jr. Day event.
If the new guidelines were implemented, Furay estimated that the department would destroy about 11.8 million body-worn videos that no longer serve a purpose.
Older videos from dashboard cameras that are still stored on magnetic videotape would also be expunged, he said. But in both cases, he added, the deletion would be done only after first checking with investigators and the department’s legal affairs section.
Several members of the civilian oversight panel, which sets LAPD policies, expressed concern that the proposed policy wasn’t clear enough.
Commissioner Rasha Gerges Shields said it lacked guardrails against the inadvertent deletion of pertinent records.
She said the proposal gave the impression that the department was “mandating ourselves to delete it as soon as the case is done.” She also asked about how videos that could be used for training purposes or other reasons would be preserved.
She asked to table a vote on the matter and told the department to report back after it had addressed the concerns.
Gerges Shields also called for LAPD officials to work with body-cam manufacturer Axon to develop a “click box” that would ensure that a video couldn’t be deleted without proper approval.
An L.A. County Superior Court judge has sanctioned the city, finding that LAPD officials probably destroyed text messages that could have provided evidence to show that police were planning to aggressively break up a 2020 demonstration outside then-Mayor Eric Garcetti’s mansion.
Since rolling out the tiny recording devices in 2015, the city has spent millions both on the cameras themselves and data storage for the digital files.
Supporters say the vast trove of footage helps provide transparency and accountability. But the recordings are rarely released to the public — and often in heavily edited ways. Critics also say it has yet to produce better behavior by police.
Most of the videos gathered by officers go unwatched, with LAPD officials saying there’s not enough manpower to review all of the clips recorded on any given shift.
The department has also struggled to keep tabs on whether officers are breaking the rules by turning off their cameras LAPD leaders said they plan to join other agencies across the country in harnessing artificial intelligence to analyze footage from traffic stops and other public encounters.
One such project is already underway, with researchers from USC and several other universities using artificial intelligence to study how officers speak to the public.
In a presentation before the commission late last year, one of the researchers said that the data-gathering portion of the study had wrapped up, and that now researchers would use the findings to “train” machine learning models for future analysis.