Science & Tech · Big Tech & Internet · 16 hrs ago
Researchers identify deepfake network targeting political issues in Canada and the U.S
McGill researchers found a network of 293 social media accounts sharing AI-made or altered videos about political issues in Canada and the United States.
The videos used real creators’ faces or existing footage, while changing voices and words to make it appear they said things they had not.
Researchers identified more than 3,000 videos with over 20 million views in three months, including content about Alberta’s referendum, elections in British Columbia and Quebec, and U.S.-Canada trade tensions.
The accounts were found on TikTok and Instagram, and the researchers said some appeared to be part of a paid operation.
The network’s source remains unknown, though the researchers said activity appeared to come from time zones in East or Southeast Asia and some accounts were registered in China.
The videos can mislead viewers and damage the reputations of people whose faces and voices are copied.
TikTok says it permanently banned accounts named in the research and removed additional accounts and content; Instagram had not responded to a request for comment.
McGill University researchers identified a network of hundreds of TikTok and Instagram accounts using deepfakes to promote content about politically sensitive issues in Canada and the United States.
The researchers found more than 3,000 videos that had drawn upwards of 20 million views in three months.
The videos focused on topics including Alberta’s referendum, provincial elections in British Columbia and Quebec, and U.S.-Canada trade tensions.
The researchers said the network’s source was unknown, though activity appeared to stem from time zones in East or Southeast Asia and some accounts were registered in China.
TikTok said it permanently banned accounts named in the report and removed additional accounts and content.
- Who
- Researchers at McGill University’s Centre for Media, Technology and Democracy identified the network.
- What
- Hundreds of TikTok and Instagram accounts used deepfakes and repurposed videos to promote content about political issues.
- When
- The videos drew more than 20 million views in the three months before the report. The report was published Thursday.
- Where
- On TikTok and Instagram, targeting Canadians and Americans.
- Why
- The videos pushed content about politically sensitive issues. The campaign’s source and purpose were not identified.
This story does not have two clearly opposing sides.
More than 3,000 videos in the network drew upwards of 20 million views.
A deepfake video reused Cole Bennett’s message urging Albertans to vote, with an AI-generated voice and mouth.
The researchers shared their findings with TikTok.
McGill researchers published their report on the network.
- Accounts identified
- 293
- Videos identified
- More than 3,000
- Views
- Upwards of 20 million in three months
- Platforms
- TikTok and Instagram
- Topics
- Alberta referendum, elections in British Columbia and Quebec, and U.S.-Canada trade tensions
3 sources
What people are saying
Words are being literally put in their mouths and shown to millions of people.
We have permanently banned the TikTok accounts named in the report and, through our own investigation, we have identified and removed additional violative content and accounts for breaking our rules on deceptive behaviours and misleading AI-Generated Content
frightening