The Republican National Committee (RNC) released an attack ad against President Joe Biden on Tuesday that featured imagery entirely generated by artificial intelligence (AI).
The ad, titled “Beat Biden,” is described as an “AI-generated look into the country’s possible future if Joe Biden is reelected in 2024.” Its release coincided with Biden officially announcing his campaign for reelection.
The footage contains manufactured images of Biden and Vice President Kamala Harris as well as numerous fictitious news broadcasts discussing imagined world events.
The ad features scenarios including China invading Taiwan, the United States’ financial system collapsing, and the U.S.-Mexico border being overrun. One clip shows AI-generated riot police on the streets of San Francisco after the city was shut down due to “escalating crime and the fentanyl crisis.”
“Who is in charge here?” a fake news anchor asks. “It feels like the train is coming off the tracks.”
A spokesperson confirmed to Axios that it is the first time that the RNC has created a video entirely from AI.
The footage is already receiving mixed reviews from those in the field of AI, including synthetic media expert Sam Gregory, who argues that the video highlights some of the challenges that arise when AI-generated imagery intersects with political campaigns.
Gregory, executive director of the human rights group Witness, specifically highlighted the importance of clear labeling on videos using AI-generated content.
“Political ads often play fast and loose with use of deceptive imagery and mis-contextualized videos and quotes, and adding in AI-generated content—image, audio, video—makes this even more complicated,” Gregory told the Daily Dot. “All these scenarios argue for strong ethical rules in campaigning and in election law around deceptive imagery, as well as better solutions to durably show how AI-generated media was made that go beyond just visible labelling.”
Giorgio Patrini, founder of the deepfake detection service SensityAI, also told the Daily Dot that the public “must embrace the new normal and quickly adapt” given how quickly AI-generated content will continue to spread. Patrini says his team is preparing to launch a new product to help organizations detect AI-generated images and text.
And while concerns have been raised about the use of AI technology during every election since 2018, synthetic media expert Henry Ajder told the Daily Dot that 2024 may “be the turning point.”
“My biggest concern remains if a realistic deepfake of a politician is not just widely believed to be real by opposition supporters, but is endorsed knowingly or otherwise by the rival candidate and their party,” Ajder said. “While this video is clearly labelled as AI-generated, it is setting a concerning precedent and sense of deepfakes being ‘fair game’ in campaigning ahead of the official presidential race that could see much more sophisticated AI-generated content being used by political parties or their supporters.”
The RNC has not revealed which technology it used to create the imagery from the ad, although Gregory opined that it could be the image-generating tool Midjourney.
In a statement on the video, the Democratic National Committee said that the GOP was forced to create fabricated imagery “because, quite simply, they can’t argue with President Biden’s results.”
