By Emmanuel Ogbodo
The announcement by President Joe Biden on Sunday that he would not seek reelection and his endorsement of Vice President Kamala Harris as the Democratic nominee has sparked a whirlwind of reactions online. However, amidst the flurry of memes and discussions, a disturbing trend has emerged: manipulated media targeting Harris.
A deepfake video purporting to show Kamala Harris delivering a bizarre and incoherent speech has recently gone viral on social media platforms such as TikTok and Twitter. The video, which features Harris speaking in front of a live audience, is a digital fabrication with AI-generated audio. Media Matters for America highlighted the video’s rapid spread on TikTok, where it accumulated millions of views before TikTok took action to remove it.
“TikTok has firm policies against harmful AI-generated content and misleadingly edited media,” a TikTok spokesperson told Mashable. “We are aggressively removing this content while partnering with fact-checkers to assess the accuracy of content on TikTok in real time.”
This particular deepfake has resurfaced online despite being debunked over a year ago. The video uses authentic footage of Harris from a 2023 speech at Howard University but digitally alters the footage and audio. In the viral clip, Harris is purportedly heard saying, “Today is today and yesterday was today yesterday. Tomorrow will be today tomorrow. So, live today so the future today will be as the past today, as it is tomorrow,” with noticeable slurring. However, Harris never made such a statement, and this segment is absent from the actual event footage.
Experts have noted digital artifacts around Harris’s mouth, indicating attempts to synchronize the fake audio with the video, as well as the absence of ambient background and crowd noises.
Despite its earlier debunking, the deepfake recently regained prominence on Elon Musk’s X platform, where a right-wing user shared it last week. The post has amassed over 3.4 million views, and X has not removed the content due to its policy guidelines, although a Community Note has been added to indicate its inauthenticity.
Unlike X, TikTok’s guidelines explicitly ban AI-generated misinformation, and the platform claims to proactively remove 98% of content that breaches its policies. Nevertheless, according to Media Matters, one viral upload of the Harris deepfake garnered over 4.1 million views before its removal. TikTok continues to work on identifying and eliminating other instances of the deepfake.
