Fake AI tornado images confuse forecasters and public, researchers warn
The Northern Tornadoes Project says convincing fake images are causing confusion and could lead to dangerous delays in official warnings.
LONDON — The director of a national tornado research group says fake, AI-generated images of tornadoes are flooding social media and complicating the work of official forecasters.
David Sills, executive director of Western University's Northern Tornadoes Project, says he first started seeing the AI images in 2025.
"Someone had taken a photograph of a storm and inserted a tornado into it," he said. "We knew it was AI because we did manage to find the original photo, so we could see exactly what was changed."
READ MORE: Mid-November remains best estimate for Toronto's first significant snowfall
READ MORE: Northwestern Ontario sees gusty winds, morning fog before sunny Monday
Sills says there has been a spike in fake images shared online in the wake of tornado warnings issued by Environment and Climate Change Canada.
In June, while an area near London, Ont., was under a tornado watch, four different AI photos circulated showing a storm far more severe than what actually occurred.
"A couple of them were actually photos of certain intersections or certain roadways, certain neighborhoods, and someone inserted a tornado into it," said Sills.
He says it can take up to two hours to investigate the more convincing fakes by reconstructing the photo's location and comparing it with radar and other storm data.
The danger, Sills says, is that the images can muddle forecasters' ability to accurately issue tornado warnings.
"When they're thinking they may issue a tornado warning, they're waiting for some evidence from a report of a funnel cloud or something on radar. And then suddenly this image pops up in the social media feeds of a tornado in the location they were considering," said Sills.
He said it is really difficult to decide if an image is real within the 10-minute target lead time for a tornado warning.
The concerns are two-fold. If forecasters post a warning based on a fake image, people may take unnecessary action. If a real tornado photo is dismissed as AI-generated, people could be put in harm's way.
Sills believes most creators of the images do not intend to cause harm.
"I think there's certain people doing it for attention or clicks, and they don't realize what they're doing is problematic to the extent that it's illegal," he said. "In other cases, it seems to be someone just sharing it with a friend as a joke, and then that friend passes it to someone else, and it quickly loses its context."
In one case, a media outlet in London picked up a fake photo and published it thinking it was real.
"Once a media organization is tricked and they put it out, it's really hard to contain," said Sills.
He says it is a growing problem that will likely get worse, and he is already seeing the first AI-generated tornado videos during Ontario storms.
Your Space Hamilton reported in September that a study found 20 per cent of House of Commons speeches may be AI-generated, sparking debate among parliamentarians.
With files from Cottage Life