"AI poses an obvious danger to millennia-long human fight to find the truth. Large language model "hallucinations," vocal deepfakes, and now increased use of video deepfakes have all had a blurring effect on facts, letting bad actors around the globe brush off even recorded events as mere "fake news." The danger is perhaps most acute in the political realm, where deepfake audio and video can make any politician say or appear to do anything."
"On Monday night, the president of the United States, a man with access to the greatest intelligence-gathering operation in the world, posted to his Truth Social account a 35-second AI-generated video filled with crude insults, racial overtones, and bizarre conspiracy theories. The video targeted two Democratic leaders who had recently been meeting with Trump over a possible agreement to fund the government;"
AI-driven technologies, including large language models and audio/video deepfakes, have blurred the boundary between real events and fabricated content, enabling bad actors to dismiss recorded evidence as "fake news." The political sphere faces acute risk because fabricated audio and video can portray any politician saying or doing anything, undermining accountability. Senior elected officials hold a special duty to model truth-seeking and responsible AI use, yet political incentives can favor weaponizing synthetic media to derail negotiations. A recently posted 35-second AI-generated clip featured crude insults, racial overtones, mariachi music, a fake mustache and a CGI sombrero, targeting two Democratic leaders.
Read at Ars Technica
Unable to calculate read time
Collection
[
|
...
]