TechnologyThe New TikTok Deepfake Hustle What's going on: You might want to pay closer attention to the videos you’re coming across on social media. While it’s no secret that misinformation and deepfakes are rampant online, it might’ve just gotten even harder to spot what’s real from what’s fake. According to an NPR exclusive, researchers found new alarming deepfake technology that can now copy real creators’ speech — stumbles, filler words, natural pauses, and all. Same script, but totally new personas, faces, different accents, no credit. And even though the posts should have been tagged as AI, the creators initially dodged moderation — as if they trained for it. What it means: Experts say we’ve entered a new phase of misinformation. AI-generated content now mimics real voices and scripts so convincingly that people scroll past without realizing they’re hearing someone else’s stolen words. And when bad actors get involved, it can be dangerous. Researchers say this could be a low-stakes test for more coordinated deception — the kind of strategy used in state-run disinformation campaigns. The UN has called for stronger tools to fight deepfakes, but platforms are already overwhelmed, and everyday users are the new front line. Related: Is Your Partner Using AI on You? There Are Some Tells (PS) |