The Deepfakes Analysis Unit (DAU) analysed a video that apparently shows United States President Donald Trump proclaiming an immediate ban on the production of Tesla electric vehicles in America. After putting the video through A.I. detection tools and getting our expert partners to weigh in, we were able to conclude that synthetic audio was used to fabricate the video.
The 58-second video in English was sent to the DAU tipline for assessment, and an Instagram link for the same video was also escalated by a fact-checking partner for analysis. The video was published on June 8, 2025 from an account with the display name of “DangerousAi”, and a display picture of a comic book character. The bio for the account mentions “I’m back” with a Ninja emoji. So far, the video has received more than 9,000 likes.
A faint watermark visible in the middle of the video frame toward the right reads “@DANGEROUSAIRETRUNS”. The handle associated with the aforementioned Instagram account and the watermark are identical, which could indicate that the suspicious video originated from this account.
Several other videos have been posted from the same account, which have been packaged in the reel format, just like the purported video of Mr. Trump. We analysed about a dozen of those reels, which mostly seem to feature international sports personalities and politicians. The nature of the content is satirical, it seems to have been created using original video footage with audio that’s not genuine.
The video being reviewed through this report shows Trump in a medium close-up, seated in what appears to be a formal setting. A blur effect extends above and below the video frame, which seems like a deliberate attempt to hide background and foreground details. His gaze seems to shift between something or someone in front of him and downwards toward the papers laid out on his desk, as if he’s reading from those.
Bold text graphics in English at the top of the video frame, visible throughout, read: “Donald Trump Bans Tesla Production in USA” followed by an emoji. Bold captions in English at the bottom of the video frame relay the content of the audio track.
A male voice recorded over Trump’s video track announces that a ban on the “production of all Teslas” will be “effective immediately”. Referring to Elon Musk, founder of Tesla and other companies including the rocket maker Space X, as “that snake Elon”, the voice alleges betrayal by him. It also states that “nobody likes Teslas anyways, unless you are a nerd”, adding that, “they catch fire and break down easily.”
The same voice refers to Trump in the first person and declares that, “I bought a Tesla a few months ago to show my support for Elon, but I have never driven it,” adding that it was “just so he can back me in the election and it worked”. It also announces that the Tesla will be listed for sale at a price of “$69”, which is “the real value of these junk cars”. The video ends in a tone of conceit as it states that, “the snake will now go bankrupt” and “he will come back begging to me in a few days again”.
On comparing the voice being attributed to Trump with that heard in his recorded speeches and interviews the similarity is evident. The accent and nasal quality of that voice and Trump’s real voice also match somewhat. However, the overall delivery sounds robotic, lacks inflection and natural pauses, which are characteristic of his style of speaking. A peculiar feature in the audio track is a fairly audible chuckle that sounds very human.
Even as the video quality is not very good, the synchronisation of Trump’s lip movements with the audio track is fairly visible. In one frame, the facial muscles on the right side of his face appear to twitch in an unnatural fashion. His teeth appear blurry with the upper set appearing and disappearing between frames. In multiple instances his right eye appears to be shut for an extended period after blinking even as the left one stays open.
We undertook a reverse image search using screenshots from the video and came across this video published on May 30, 2025 from the official YouTube channel of The Wall Street Journal, an American business news publication. Trump’s clothes in this video and the one we are analysing are identical and his body language also looks the same.
However, the background and foreground in the two videos are not identical as barely a few objects and details from the source video have been captured in the doctored video, which carries more zoomed-in frames of Trump. The other difference between the two videos is that the objects seem to be positioned in the opposite direction.
The American flag pin on Trump’s lapel appears on the right side in the source video but in the doctored video it’s on the left. A photo placed in a frame, and part of the American flag, U.S. Marine Corps flag visible on the right behind Trump’s chair in the doctored video, appear on the left in the source video. This establishes that the doctored video has used a mirrored version of the visuals taken from the source video.
The audio tracks in the two videos are different though both are in English. Trump does not announce any ban on Teslas in the source video, which features other subjects as well, including Mr. Musk, but none of them are part of the doctored video. Static text graphics in English appear in the source video at the bottom of the video frame, which are different from those in the doctored video. The logo of the Wall Street Journal can be seen in the top right of the video frame with the words “LIVE” below it.
To discern the extent of A.I. manipulation in the video under review, we put it through A.I. detection tools.
The voice tool of Hiya, a company that specialises in artificial intelligence solutions for voice safety, indicated that there is a 62 percent probability of the audio track in the video having been generated or modified using A.I.
Hive AI’s deepfake video detection tool highlighted several markers of A.I. manipulation throughout the video track. Their audio detection tool indicated that most of the audio track is A.I.-generated but for a 10-second segment.
We also ran the audio track from the video through Deepfake-O-Meter, an open platform developed by Media Forensics Lab (MDFL) at UB for detection of A.I.-generated image, video, and audio. The tool provides a selection of classifiers that can be used to analyse media files.
We chose six audio detectors, out of which three indicated that it was highly likely that the audio track in the video is A.I.-generated. AASIST (2021) and RawNet2 (2021) are designed to detect audio impersonations, voice clones, replay attacks, and other forms of audio spoofs. The Linear Frequency Cepstral Coefficient (LFCC) - Light Convolutional Neural Network (LCNN) 2021 model classifies genuine versus synthetic speech to detect audio deepfakes.
RawNet3 (2023) allows for nuanced detection of synthetic audio while AVSRDD (2025) is an audio-visual speech recognition based deepfake detection method. Whisper (2023) is designed to analyse synthetic human voices.
To check for elements of A.I. in the audio we ran it through the A.I. speech classifier of ElevenLabs, a company specialising in voice A.I. research and deployment. The results that returned indicated that it was “very unlikely” that the audio track used in the video was generated using their platform.
We reached out to ElevenLabs for a comment on the analysis. They told us that based on technical signals analysed by them they were able to confirm that the audio track in the video is A.I.-generated. They added that they have taken action against the individuals who misused their tools to hold them accountable.
We also escalated the video to our partner GetReal Security, co-founded by Dr. Hany Farid and his team, they specialise in digital forensics and A.I. detection. They stated that there is evidence suggesting that the video has been manipulated.
They used multiple digital analysis models, including spectrogram analysis, contextual analysis, and human speech analysis on the video’s audio track and all pointed to synthetic generation.
They stated that the audio also exhibits unnatural “reverberation truncation” and lacks background room tone. They compared the audio with recorded authentic audio tracks of Trump speaking from the same position in the same room and found no such truncation but did find background room tone.
Echoing our observations, the team noted that it is odd that the video is horizontally flipped or mirrored and that it is closely cropped around the subject, removing contextual clues that could be used to more easily identify the source video. They too pointed to the original visuals being from a press conference held on May 30, 2025 in the Oval Office as we noted above.
Based on our findings and analyses from experts, we can conclude that original footage featuring Trump was mirrored and used with synthetic audio to fabricate the video.
(Written by Debraj Sarkar and Rahul Adhikari, edited by Pamposh Raina.)
Kindly Note: The manipulated video/audio files that we receive on our tipline are not embedded in our assessment reports because we do not intend to contribute to their virality.
You can read below the fact-checks related to this piece published by our partners:
Fact Check: This Video Of Trump Announcing Ban On Tesla Is A DeepFake
Viral Video Of Trump Announcing Ban On Tesla Production Is A Deepfake
Trump Vs Musk: US President Bans Tesla Production? No, Viral Video Is Deepfake
AI-Manipulated Video of Trump Banning Tesla Production in the US Viral as Real
.jpg)
