How to spot a deepfake —and prevent it from causing political chaos

The “fake news” of 2016 is so passé. This year, fake video and audio clips generated by artificial intelligence (AI) are the new looming threat that could sway voters in the upcoming U.S. presidential election. Last week, voters in New Hampshire’s presidential primary received a robocall in which President Joe Biden seemed to be discouraging them from participating in the election. “Voting this Tuesday only enables the Republicans in their quest to elect [former President] Donald Trump again,” the recording said. Only the call wasn’t actually recorded by Biden—it was a deepfake, a form of generative AI in which algorithms collate clips of a person’s face or voice, learn from them, and impersonate the subject saying things they never did. Experts are alarmed at how convincing it seemed. “It has a lot of dangerous things combined in one: disinformation, AI-generated voices, impersonating the president, and essentially discouraging voting, which is an illegal activity,” says data scientist Matt Groh of Northwestern University. Deepfakes are cheaper and easier to produce than ever, and we’re likely to see many more during the election season. Science talked with several experts about the dangers of AI-generated content, and why we struggle so much to recognize it. Are deepfakes really any worse than old-fashioned fake news? False narratives on social media have certainly caused problems for years. Researcher...
Source: ScienceNOW - Category: Science Source Type: news