The internet is awash in deepfakes — audio, pictures or video made using artificial intelligence tools in which people appear to do or say things they didn’t, be somewhere they weren’t, or that change their appearance. Some involve nudification, where photos are altered to depict someone unclothed. Other deepfakes are deployed to scam consumers, or to damage the reputations of politicians and other people in the public eye.
Advances in AI mean it takes just a few taps on a keyboard to conjure up a realistic deepfake. Alarmed governments are trying to fight back, but it’s been a losing battle. Fraud attempts using deepfakes have grown more than 20-fold in the past three years, according to data from identity verification company Signicat.
Where have deepfakes been in the news?
On July 20, US President Donald Trump circulated a deepfake video appearing to show former President Barack Obama being arrested inside the Oval Office. The video had appeared on TikTok before being reposted on Trump’s Truth Social account, without any comment.
Imposters have used voice cloning tools to pose as prominent politicians, with recent cases involving Trump’s predecessor Joe Biden and Secretary of State Marco Rubio. During the 2024 US presidential election, Elon Musk shared a deepfake campaign video featuring Democratic candidate Kamala Harris without labeling it as misleading. The video’s AI-manipulated voice appeared to show her calling Biden senile and declaring that she didn’t “know the first thing about running the country.” The video garnered tens of millions of views. In response, California Governor Gavin Newsom vowed to ban digitally altered political deepfakes, and signed a law doing so in September.