Police in Thunder Bay, Ontario, are piloting a new video analysis tool that will dramatically reduce the amount of time needed to sift through surveillance footage. The city has already installed 13 high-resolution cameras as part of its Eye on the Street surveillance program, and is boosting that number to 20 as part of the pilot.
The footage captured with those cameras will be analyzed with BriefCam video analytics software. BriefCam has face and license plate recognition capabilities, and as well as search tools that allow Thunder Bay Police Services (TBPS) officers to set more specific parameters for a search. For example, the software can find every man wearing a white shirt and black pants in a given period, or track a red vehicle that was moving in a specific direction. The software can identify 28 different classes and attributes, and sort information based on factors like size, color, speed, direction, and loitering time.
According to TBPS, BriefCam needs less than an hour to sort through a day’s worth of footage and return a potential lead, making it much faster than manual review. TBPS believes that those savings will be critical during time-sensitive investigations.
Of course, any automated surveillance system raises major privacy concerns, and TBPS indicated that it is taking those concerns seriously. The department indicated that the Police Services Board will pass a governance policy that sets clear guidelines for the use of the technology, and claimed that it is working with Ontario’s Information and Privacy Commissioner to ensure that that policy adheres to technical and ethical best practices. TBPS also stated that none of the footage captured during the pilot will be stored permanently on its servers.
The initial BriefCam pilot will run for a six-month period. Members of the public will have opportunities to weigh in on the program while it is in progress, while TBPS will be expected to submit a report that details the impact of the technology. That report should address any concerns about bias or accuracy that arise during the pilot.
Canada’s federal government is currently trying to defend the RCMP from a class action lawsuit that stems from the agency’s unlawful use of Clearview AI’s facial recognition technology. The RCMP in British Columbia have also been accused of misleading the public about their use of facial recognition.
June 7, 2022 – by Eric Weiss