An avalanche warning was issued for the upper reaches of India’s northern state of Uttarakhand on January 27, but an AI-generated video purported to show snow sliding down a mountainside and burying onlookers was shared on social media. This clip was previously shared by an account indicating it was created using the text-to-video conversion AI model Sora, and contains discrepancies indicating synthetic content.
“Avalanche in Uttarakhand,” reads the overlapping text and caption of an Instagram video shared on February 6, 2026.
The video shows an avalanche rushing down the mountain and crashing into people and vehicles on a narrow street.
Screenshot of a false post taken on February 10, 2026. Red X and AI symbol added by AFP
The same footage was also posted in similar posts on Facebook and Instagram after heavy snowfall hit the northern Indian state of Uttarakhand and an avalanche warning was issued (archive link).
An avalanche struck a tourist destination in Jammu and Kashmir on the night of January 27, with dramatic footage shared by Press Trust of India (archive link) showing snow crashing and engulfing the site. Officials said there was no loss of life.
India’s Himalayan region is prone to avalanches and irregular winter rains, but circulating videos do not show the avalanche that hit Uttarakhand (archived links).
a reverse image search A Google search using keyframes from the accidentally shared video revealed that the same clip had previously been shared on Instagram on January 30th by an account that regularly shares AI-generated content (archived link).
The caption for the Jan. 30 clip includes a reference to the text video AI model Sora.
Comparison of accidentally shared video (L) and screenshot of Instagram clip from January 30th
A detailed analysis of the mistakenly shared videos also revealed visual inconsistencies, such as incomprehensible text appearing on store signs and people on the street watching the avalanche but not reacting.
Screenshot of visual discrepancies in accidentally shared video highlighted by AFP
We used the Hive Moderation AI video detection tool to analyze the video and found it “may contain AI-generated or deepfake content” (archive link).
Hive moderation analysis screenshot
AFP previously Error exposed Other false claims about natural disasters using AI-generated footage.
