AI-Faked Videos and Fake Canadiens Ties: The Disinformation Wave After the White House Correspondents' Dinner Shooting

Since Cole Tomas Allen was charged with attempted assassination following the April 25 shooting at the Washington Hilton, the online information environment around the case has been a mess of AI-generated fabrications.

Allen, a 31-year-old tutor and amateur video game developer from Torrance, California, is accused of breaching security at the White House Correspondents' Dinner carrying two firearms and knives, according to CP24. He faces charges of attempted assassination plus two firearms counts. None of that is disputed.

What followed online was a different story entirely.

The AI-Enhanced Security Footage

In the hours after the shooting, President Trump posted low-quality security footage to Truth Social showing a person running through a checkpoint at the Washington Hilton as Secret Service agents pursued with guns drawn. That clip was the seed of the first disinformation wave.

X user Seth Weathers ran the footage through an AI image-enhancement tool and reshared it. Weathers acknowledged in his own post that he had enhanced the footage and that "the AI made up some things to fill in the gaps," according to Poynter.

That caveat did not travel with the video. An April 26 Facebook post circulated the AI-processed version with the caption describing it as "unedited raw security footage," which was false. Conservative commentator Benny Johnson shared the same clip on X, where it accumulated more than 2 million views. Johnson later noted in a follow-up post that the video had been AI-enhanced, but by then the original spread had already happened, according to Poynter.

Poynter and its PolitiFact unit rated the claim that the footage was unedited and raw as False.

AI upscaling tools fill missing visual data with plausible-looking pixels. They do not recover real information. Any forensic conclusion drawn from the enhanced version is worthless and potentially wrong.

The Fake Montreal Canadiens Connection

A separate disinformation operation manufactured a different kind of story. Multiple Facebook posts, many beginning with the word "BREAKING," circulated AI-altered images of Allen in Montreal Canadiens gear, including a lanyard with the Canadiens logo attached to a badge reading "MONTREAL CANADIENS STAFF," according to CP24.

Chantal Machabée, vice-president of communications for the Canadiens, confirmed to CTV News in a direct statement: "Cole Allen never worked for the Canadiens." The images were fabricated.

The same operation also produced posts linking Allen to the Minnesota Vikings, the Detroit Lions, and other sports organizations. The pattern was identical across all of them: AI-altered images, fake sports-staff credentials, no factual basis.

One of the accounts spreading the Canadiens version presented as the official page of Olympic hockey players Monique and Jocelyne Lamoureux. It had nearly 69,000 followers and a Meta-issued blue verification badge, lending it surface credibility. CTV News found the account is managed by five accounts based in Vietnam and one in Sierra Leone. It has been linked to more than 450 inactive ads removed by Meta for policy violations, many targeted at European audiences.

The motive, according to Mathieu Lavigne, analytic lead at the Media Ecosystem Observatory at McGill University, was straightforward: money. The posts linked to ad-saturated fake news websites designed to generate revenue from clicks.

The Stronger Concern

Some users and commentators argued that the original security footage Trump shared was too low-quality to allow the public to independently verify the government's account of events, and that in the absence of clear footage, people naturally sought better visual information. That concern is reasonable. Transparent, high-quality public disclosure of evidence in cases involving alleged assassination attempts serves the public interest.

That concern does not, however, validate the AI-enhanced clips as evidence. The problem is not that people wanted better information. The problem is that AI tools generated plausible-looking fabrications, those fabrications were labeled as authentic, and millions of people shared them without the correction.

The Unresolved Platform Question

The disinformation operations that ran alongside the initial news cycle have been documented and debunked. The unresolved question is whether Meta will take enforcement action against the Vietnam- and Sierra Leone-managed account that spreads AI-fabricated disinformation under the identity of American Olympic athletes — accounts that hold Meta verification badges and have run hundreds of ads violating the company's own standards.

As Lavigne put it: "I think that Facebook needs to seriously evaluate why it is the case that users are seeing a lot of all this AI slop from unknown accounts and AI-generated content in their feed, and what they can do to make sure that the platform is safe for users."