发表在PLOS One上的一项新研究显示，AI聊天机器人（特别是GPT-4 Turbo）能够令人信服地模仿公众人物，生成的回复被认为比真人更真实、更连贯。研究人员提示AI模仿了来自英国的112位公众人物，使用了他们维基百科传记的信息以及BBC节目“Question Time”的格式。参与者认为AI生成的回复优于真实辩手的回答，这凸显了AI驱动的虚假信息带来的重大风险，尤其是在政治背景下。 AI
影响 AI令人信服地模仿公众人物的能力对虚假信息构成了重大威胁，尤其是在政治领域。
排序理由 发表在PLOS One上的研究，详细介绍了AI的能力和风险。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
Researchers discovered that people found AI impersonators to be more authentic, coherent, and relevant than the real politicians, raising alarm bells around the potential for public deception.
Scientists Asked # AI to Impersonate 112 Public Figures. What Happened Next Is a ‘Dire’ Warning https://www. 404media.co/untitled-28/ # privacy # deepfake
Scientists used AI to impersonate 112 public figures including politicians and journalists. A new study found that AI-generated responses were perceived as more authentic, coherent and relevant than the real thing. Researchers warn this poses a dire risk of disinformation ahead o…