News / Winnipeg

'Using technology to fight technology:' U of M researchers aim to prevent child porn distribution

Two computer engineering students are working with a B.C. company to create cutting-edge artificial intelligence software aimed at intercepting child porn.

(Left to right) Binglin Li and Mehrdad Hosseinzadeh are PhD students working on a child pornography detection project, photographed February 1, 2017 at the University of Manitoba.

David Lipnowski / Metro Order this photo

(Left to right) Binglin Li and Mehrdad Hosseinzadeh are PhD students working on a child pornography detection project, photographed February 1, 2017 at the University of Manitoba.

Two researchers from the University of Manitoba are developing technology aimed at blocking images of child sexual abuse before they have a chance to infiltrate the Internet.

Computer science students Mehrdad Hosseinzadeh and Binglin Li are working with Two Hat Security Ltd. out of Kelowna, BC to create cutting-edge artificial intelligence software to accurately identify and intercept pictures and prevent distribution.

“It’s using technology to fight technology,” said Brad Leitch, Two Hat Security’s lead on what has been dubbed “Project Cease.”

“When you look at how much new content is being uploaded to the Internet, the statistics are scary,” Leitch said. “With video cameras virtually on the hips of everyone in the form of phones, it has become even more of a challenge for law enforcement and child protection agencies to keep up.”

“Project Cease” is supported by Mitacs, a national, not-for-profit organization that designs and delivers research and training programs in Canada using government funding.

Hosseinzadeh said he and his research partner are thrilled to be involved with such a “groundbreaking” project. He also admits there are some known hurdles they must face.

“It’s going to be a little bit tricky because it’s against the law for us to view the actual content that we are training computers to identify,” he said. Further, in the developers are not allowed to access or store any such content. They're merely developing the algorithm, and eventually the RCMP directly apply the algorithm on their content.

“We are confident we can produce this very important software tool, but we are sort of doing it with blinders on,” added Hosseinzadeh.

The end goal is to have this software tool proactively stop someone from uploading child sexual abuse material. For example, it will be possible for the software to send a warning that the image they are about to upload from their phone, tablet or computer is inappropriate or illegal.

Lloyd Richardson, IT director for the Canadian Centre for Child Protection, said this technology will be a great compliment to a new tool they recently launched called “Project Arachnid.” The automatic web crawler detects images and videos by analyzing digital information.

“The more technology we have available to combat child sexual abuse, the better,” Richardson said. “And to be able to find images before they can even make it on the Internet is a huge win.”

Correction: An earlier version of this story incorrectly stated that Mehrdad Hosseinzadeh and Binglin Li were computer engineering students. This story has also been updated to clarify that developers can't view, access or store any content depicting child pornography.

More on Metronews.ca