A network of 293 TikTok and Instagram accounts published 3,201 videos that accumulated 22.1 million plays between 7 August and 4 November 2026, researchers at McGill University's Centre for Media, Technology and Democracy report. The videos use manipulated footage of real creators to deliver political commentary, including fabricated trade-war claims.
The total counts plays across the identified network, not individual viewers or exclusively false trade stories. TikTok subsequently said it banned the accounts named in the research, which was released on 12 November.
Real Footage Carries Manufactured Speech
The accounts copy existing creator videos, add synthetic narration and alter the mouth to match the new speech. Much of the original footage remains, making a real person appear to deliver words they never said.
In one documented example, footage of American creator Ashley Hetherington discussing dating reappeared with narration about Canadian jobs, tariffs and Alberta's referendum. Her head, hair and eyes followed the original recording, while her lips were digitally redrawn.
Researchers connected the accounts through shared video formats, captions, scripts and even repeated typos. Those patterns linked apparently separate narrators to a common production operation. Almost every identified account relied on footage of a single real person, giving its changing scripts a consistent face.
'Words are being literally put in their mouths and shown to millions of people', researcher Aengus Bridgman told The Globe and Mail.
One Invented Story Spread Across 129 Accounts
The network's most-played story fabricated a confrontation between US President Donald Trump and Canadian Prime Minister Mark Carney. It claimed Trump issued a 48-hour ultimatum for Canada to remove retaliatory tariffs and Carney responded by cancelling $8 billion in US defence contracts.
Neither event happened, according to the report. Yet the script appeared in 161 videos across 129 accounts, collecting 5.1 million plays between them. That figure represents repeated versions of the same invented story, rather than one upload attracting the entire audience.
Trade-war material accounted for 71% of the network's plays, but the accounts did not consistently support one side. Narrators attacked both governments and sometimes presented themselves as Canadian in one video and American in another.
The repetition gave the fabrication multiple apparent speakers, while the reused footage concealed that the original creators had not delivered those statements.
TikTok Reports Bans After Researchers Share Findings
TikTok's AI label appeared on approximately 4% of the checked network videos built on real creators' footage. Researchers found labels on 26 of 709 such videos; this is a finding about their sample, not TikTok's overall detection rate.
The Globe and Mail reports that researchers shared their findings with TikTok on Wednesday evening. The platform then announced action against the identified network.
'We have permanently banned the TikTok accounts named in the report', spokesperson Danielle Morgan said. The company also said its own investigation identified additional accounts and content that it removed for violating rules on deceptive behaviour and misleading AI-generated content.
The supplied reporting contains no equivalent removal announcement from Instagram, which hosted 36 identified accounts. TikTok's response therefore does not establish that the entire cross-platform operation has disappeared.
Researchers have not identified who runs or funds the network, or established its purpose. The removal of identified accounts leaves those questions unanswered.
