THUNDER BAY — A new search tool being rolled out this year by Thunder Bay Police holds the potential to significantly accelerate the pace of certain investigations.
BriefCam video analytics software uses artificial intelligence to detect and extract people and objects of interest from multiple sources with speed and precision.
The software was acquired recently as part of the modernization and expansion of the City of Thunder Bay's Eye on the Street surveillance program.
TBPS has started a six-month pilot project, but Director of Communications and Technology Chris Adams is already confident it will become an important investigative tool.
"Our investigators will be able to search vast amounts of video much faster than having to manually go, essentially second by second, through the various videos. For Eye on the Street video, that's pretty critical," Adams said in an interview.
The surveillance program has been expanded from 13 cameras to 20 high-resolution cameras at various locations around the city.
Adams said the new software can be used in various investigations, "Everything from the things that you would expect to be captured on street video, such as assaults, but we can also use it more effectively for things like missing person investigations and abductions."
Video analytics have become much more important now because of the sheer volume of material that's available.
On its website, the company that markets BriefCam promises unmatched detection and extraction accuracy by quickly searching and filtering objects and events of interest by men, women, children, vehicles and lighting changes, using 28 different classes and attributes.
The system also utilizes face recognition, licence plate recognition, appearance similarity, colour, size, speed, path direction and dwell time, "providing an ever increasing and powerful set of distinct search combinations," the company states.
As an example, Adams explained how it can be programmed to search for somebody wearing a white shirt and black pants, and can expeditiously analyze hours of video to pinpoint all instances where a person matching that description appeared on one of the Eye on the Street cameras.
"It's also very useful for looking at vehicles, whether cars, trucks, or bicycles. It can determine direction. It does a lot of very powerful things using A.I. to speed that process up. If you were looking at approximately 24 hours of video, the processing time can be less than an hour."
The ability to shrink search times so dramatically could be instrumental in solving critical cases such as missing persons, he said, in contrast to "Having an individual sitting in front of a screen and watching video in real time, when you're talking about multiple video sources. That can be a tremendous amount of time and resources."
Adams said there's also considerable potential for finding things that may be missed in manual scanning.
Facial recognition technology raises red flags
The use of facial recognition technology has raised red flags with civil rights groups and with privacy commissioners all across Canada.
Last month, Ontario Information and Privacy Commissioner Patricia Kosseim called on the House of Commons to enact laws that specifically govern its use by police.
Kosseim said "Clear guardrails with the force of law are necessary to ensure police agencies can confidently make appropriate use of facial recognition technology, grounded in a transparent framework capable of earning the public's enduring trust."
She recommended several key elements including quality and accuracy controls to avoid false positives and reduce the potential for bias against individuals, groups and communities, as well as clear limits for how long images will be retained if they do not register a match.
Kosseim also said police must adopt strong accountability measures, including conducting a privacy impact assessment before deploying the technology.
Adams acknowledged these kinds of concerns, saying they are being expressed all over North America as the technology goes into more widespread use, but said TBPS is consulting closely with the Information and Privacy Commissioner.
"We are very cognizant of that...You could potentially misuse it if you didn't understand the limitations. That's the key here...understanding where the potential traps are that could see it misused."
He said that although police need modern investigative tools to do their job, it's important that nobody is marginalized in the process.
"So the incidents in which it would be used would have a lot of caveats, a lot of restrictions attached to protect that."
According to Adams, TBPS will use the new technology under a governance policy approved by the Thunder Bay Police Services Board, and will report back to the board on how it works and what the implications are.
"They set basically what the expectations are, and we meet those expectations through our practices."
During the current pilot phase, none of the video being processed during the test will be used or retained.
Adams said there will be an opportunity for public input in the coming months, and that TBPS is also working with other police services that are already using BriefCam.