Fake AI-generated video falsely claims Tim Walz assaulted a gay student

A fake, AI-generated video which went viral falsely claims that Kamala Harris’s running mate Tim Walz sexually assaulted a gay student he taught in the 90s.

The post, which has been removed from X/Twitter but remains available to view in an archived form, alleges a former student of Walz named Matthew Metro had come forward to reveal the vice-presidential hopeful assaulted him.

In the video, which was identified by AI-detector TrueMedia as highly likely to have been generated, the fake Metro recounted reaching out to Walz for guidance during a difficult period in his senior year and said he was groped and kissed by him.

“Breaking: Tim Walz’s former student, Matthew Metro, drops a shocking allegation, claims Walz s*xually assaulted him in 1997 while Walz was his teacher at Mankato West High School. Metro was a senior at the time. If this is true, it’s a political earthquake,” the caption read.

Metro is a real person who attended that school, where Walz was a teacher and football coach but has said he never interacted with the Minnesota governor during his final year, a time he described as a “breeze”.

Speaking to The Washington Post, the real Metro, who now lives in Hilo, Hawaii, said: “It’s obviously not me: the teeth are different, the hair is different, the eyes are different, the nose is different. I don’t know where they’re getting this from.”

The fake Metro said his parents got divorced and he was not open about his sexuality. However, the real person, who is gay, insisted he was out in high school and his parents are still married.

“I was completely out in high school,” he said.

Now 45, the former student has been left angered by the AI content, labelling it an “invasion of my privacy and my personal life”.

I’ve analyzed the first 30 seconds of the video being shared on Twitter (X), allegedly featuring someone claiming to be Matthew Metro. Immediately, there are several indicators pointing toward a deepfake.

I’ll play the first 25 seconds of the original video.. then I’ll show you… pic.twitter.com/dfWs1UVYSU

— Agent Self FBI (@FraudWatch89) October 17, 2024

The AI video was watched millions of times and was re-shared by right-wing accounts which support the Republican ticket of Donald Trump and JD Vance.

BBC disinformation journalist Shayan Sardarizadeh took to X to say: “That video is a deepfake. If it’s not instantly obvious to most people that it is a deepfake, then we’re soon going to be in serious trouble.”

Other users pointed out the tell-tale signs of a deepfake, including unusual facial movements, unchanging speech patterns and inconsistent physical features.

One user, who was linked in the Community Notes under the original video, said: “First, pay attention to how this person speaks. The speech pattern strongly suggests it’s an AI-generated prompt. For example, listen to how he says “sexual assault” three times within the first 30 seconds.

“Each time, it’s said with the exact same pitch, tone and inflection. That level of consistency is unnatural for a human. There’s no natural progression in human language. This is scripted and not someone recounting a memory from emotional experience.

“If you try repeating a phrase multiple times, the pitch and tone will always vary slightly. When you run an audio analysis on this video, the results show identical pitch and tone across all three instances of the same words: ‘sexual assault’. Any freeware Audi wave editor will demonstrate this, which is a strong indicator that the audio has been generated by a machine.

“Secondly, visual cues further suggest this is a deepfake. The facial movements are off: there’s inconsistent muscle movement, drooping of the eyes, irregular eyebrow shifts, and even the disappearance and reappearance of a tooth throughout the video. These are all signs that the video has been AI-generated.”

Share your thoughts! Let us know in the comments below, and remember to keep the conversation respectful.