On Friday, X owner Elon Musk shared a deepfake of Vice President Kamala Harris on his X account, potentially violating his platform's policies on synthetic content.
Why this matters. This incident raises concerns about AI-generated deepfakes and their ability to spread disinformation, especially with the approaching U.S. presidential election.
Some perspective. Deepfakes, which use AI to create convincing yet fake audio or video of real people, pose a growing threat to public trust in information, the spread of hoaxes, and election integrity.
In 2018, a Barack Obama deepfake gave us one of the first indications that they would become easier and quicker to produce, with significant consequences for disinformation.
In detail:
- The synthetic video features Harris’ voice but with fake words dubbed over her actual speech.
- In the video, Harris is made to say that Biden is senile, that they’re puppets of the government, and that she was chosen as the Democratic nominee because she’s “the ultimate diversity hire.”
- The fake video also includes the statement, “I’m both a woman and a person of color, so if you criticize anything I say, you’re both sexist and racist.”
- Originally posted as a parody, Musk shared it without any warning, only commenting, “This is amazing.”
- The video has been viewed more than 130 million times and counting.
- The images used in the video are from a real Harris campaign ad.
Reading between the lines. X’s terms of service prohibit sharing manipulated material that could mislead or confuse people, and places full responsibility on the disseminator.
Musk’s post fits this prohibition, but it remains unremoved and untagged.
More context.
- Musk has endorsed Donald Trump in the 2024 presidential election.
- X allows memes and satire as long as they don’t lead to confusion.
- The fake video is good enough to assume that many people will take it as authentic.
The other side. In her campaign, Harris has criticized the “fake, manipulated lies” of her rival Trump, but also those of Musk.
Questions now remain as to how the use of AI to create manipulated but convincing videos on a massive scale will be addressed.
Bottom line. As AI tools become faster, more efficient, and more accessible, the potential for disinformation campaigns and election interference grows.
Sooner or later, there will presumably be much tougher legislation against the dissemination of such material masquerading as real.
This article was written by Javier Lacort and originally published in Spanish on Xataka.
Image | Xataka (using Midjourney)
Related | Deepfakes From Met Gala Prove We Have a Problem. Even Katy Perry’s Mom Thought They Were Real
View 0 comments