Edmonton police first in world to test Axon facial recognition body worn video cameras

December 3, 2025
6 min read
Edmonton police first in world to test Axon facial recognition body worn video cameras
The Edmonton Police Service is the first in the world to test facial recognition body worn video cameras from Axon Enterprise Inc., an American company that develops weapons and technology products for law enforcement, military and civilians.

It’s not from Cyberpunk 2077 or RoboCop.

It’s new camera technology powered by machine learning that could let police identify people deemed dangerous, and those with arrest warrants, in real time.

Edmonton Police Service (EPS) is the first in the world to test facial recognition body worn video cameras from Axon Enterprise Inc., an American company that develops weapons and technology products for law enforcement, military and civilians.

The program will launch Wednesday, with up to 50 Edmonton police officers using facial recognition-enabled body worn video cameras on their shifts for the remainder of the month.

EPS Acting Supt. Kurt Martin told reporters at a Tuesday press conference the end goal is for the technology to enhance officer and public safety.

“It’s about empowering our officers to be informed about safety risks to themselves and to the public, as well as giving them a tool to help them in real-world operations in near real-time,” he said.

Standard body worn video cameras have been used by EPS since September 2024 after the Government of Alberta mandated all police agencies adopt them.

Martin said the original cameras improved evidence collection, documentation and led to successful prosecutions. He also said the cameras have helped with transparency and building public trust.

Now, he said, EPS is taking the next step with facial recognition.

Cameras identify suspects with outstanding warrants

Martin added the pilot will test if the cameras can work with EPS mugshots and correctly identify people deemed potentially dangerous based on previous police interactions, and those with outstanding warrants for serious crimes, such as murder, aggravated assault and robbery.

During the test, officers will be using the cameras in a “silent mode,” where they will be recording and potentially identifying people, but won’t be informing the officer.

A team of officers with expertise in the area will later review footage to evaluate whether the facial recognition software is working as intended.

In a real-use case scenario, Martin told reporters that any possible resemblance notifications will be double-checked by trained officers for accuracy.

“I want to make it clear that this facial recognition technology will not replace the human component of investigative work,” Martin said.

After the test period, any still images used for facial recognition will be deleted, but the original video will be kept according to EPS rules. All collected data will be encrypted and stored in Canada, Martin said.

Machine-learning and facial recognition

The company providing the cameras, Axon, is a partner of EPS and will be providing the technology for the test, free of charge. The only extra costs, Martin said, will be related to staffing officers reviewing resemblance notifications.

Axon is a popular manufacturer of standard body worn video cameras and is known for inventing the Taser and has also considered adopting facial recognition for body-worn cameras in the past.

The company released a report in 2019 from its AI ethics board concluding that facial recognition technology at the time was not yet reliable enough to justify body worn video cameras due to poor performance when it came to race, ethnicity, gender and other groups.

Ann-Li Cooke, Axon Enterprise Inc.’s director of responsible AI, told reporters there has since been more research and training done to fill gaps.

Cooke said Axon evaluated multiple machine learning models and found race was not the limiting factor; it was skin tone. Varying conditions such as distance and dim lighting can make it harder to match and detect people with darker skin, Cooke said.

She added that the proof of concept test will include guardrails where the cameras will be restricted to only detect people within four metres and will only be used in good lighting.

When asked about the specific machine learning model that powers Axon’s facial recognition, Cooke said the company uses third-party models and is still evaluating them. She said Axon will be using just one model for EPS’s proof of concept testing.

Cooke wouldn’t disclose which model was being used.

AI law expert worried about privacy, bias

Gideon Christian is an associate law professor at the University of Calgary and specializes in AI and law.

He said he was shocked that EPS is choosing to be the testing ground for the new cameras.

“Body-worn cameras were originally defined as a tool for police transparency and accountability,” Christian said. “This tool is basically now being thrown from a tool for police accountability and transparency to a tool for mass surveillance of members of the public.”

EPS said it has submitted a privacy impact assessment to Alberta’s information and privacy commissioner to ensure the facial recognition proof of concept is fair and respects people’s privacy and the law.

Martin said he’s not sure when EPS can expect results.

While EPS says it has guard rails and oversight for the proof-of-concept test, Christian said he’s worried about what might happen if the technology is fully deployed.

When it comes to privacy concerns, Christian said that people could be recorded and identified without their consent. He also said there is a potential civil liberty issue and that the cameras could have a chilling effect on free speech and association.

In some cases, such as a protest, Christian said some people could be worried about attending if they know police are wearing these cameras and could record and identify them.

He also mentioned that facial recognition models have known problems identifying people with darker skin tones and could lead to problems.

“Can you imagine being a darker skinned individual, and with high error rate in this tool, you are falsely matched to a suspect?”

When testing is complete, the Edmonton Police Commission and the chief’s committee will review results and consider whether to move forward with further testing in 2026.

lnewbigging@postmedia.com

Related


Bookmark our website and support our journalism:
Don’t miss the news you need to know — add EdmontonJournal.com and EdmontonSun.com to your bookmarks and sign up for our newsletters here.

You can also support our journalism by becoming a digital subscriber. Subscribers gain unlimited access to The Edmonton Journal, Edmonton Sun, National Post and 13 other Canadian news sites. Support us by subscribing today: The Edmonton Journal | The Edmonton Sun.

Share this article