Fake AI-generated audio files of Germany’s most popular news program, Tagesschau, are currently spreading on social media and have even been played at a demonstration.
The fake audios contain supposed apologies for alleged lies in reporting on topics such as the war in Ukraine and the coronavirus pandemic, as well as the “denunciation” of demonstrators. They were played at a demonstration in Kiel, Germany, and distributed via conspiratorial Telegram channels, Facebook, and TikTok.
Marcus Bornheim, editor-in-chief of ARD-aktuell, clarifies that the files are fake. The voices in the audio files sound mechanical and have false intonations.
“The positions and strategies represented in the AI-generated audio files are capable of damaging the free and democratic basic order,” says editor-in-chief Bornheim. The good reputation of the Tagesschau is being exploited “to spread targeted disinformation,” Bornheim adds.
News anchors are easy targets for deepfakes
Just recently, German news anchors Christian Sievers and André Schünke were victims of a deepfake scam in which their digital copies were used in fraudulent promotional videos on Facebook.
The videos promoted AI-powered investment platforms and promised high returns. The cloned voices of the news anchors sounded deceptively real, and the synchronized lip movements were convincing enough for those quickly scrolling through social media.
News anchors and shows are ideal targets for such scams because they are well known to the public, and there is a wealth of video and audio material available online that can be used to train deepfake AI systems.
Experts warn of the dangers of such fakes, which can undermine the credibility of news anchors and media outlets. AI-generated fakes are likely to increase in the future as they become easier to produce. Their quality is also improving.