OpenAI的新 Sora 2 App从文字中产生AI视频, 引起人们对错误信息、版权和深层假象的担心, OpenAI's new Sora 2 app generates AI videos from text, sparking concerns over misinformation, copyright, and deepfakes despite safeguards.

OpenAI已推出Sora 2, 一个社交媒体应用程序, 从文字提示中产生10秒的AI视频, OpenAI has launched Sora 2, a social media app that generates 10-second AI videos from text prompts, resembling TikTok content. 它允许制作真人像的视频(经许可), 包括隐私管制和AI水印, 并禁止有害内容。 It allows creation of videos with real people’s likenesses (with permission), includes privacy controls and AI watermarks, and bans harmful content. 然而,测试发现,它仍然可以产生误导性视频,包括虚假的历史主张和暴力描述,并产生带有商标字符的内容,引起版权和错误信息方面的关切。 However, testing found it could still produce misleading videos, including fake historical claims and depictions of violence, and generate content with trademarked characters, raising copyright and misinformation concerns. 尽管有保障措施,但应用程序的默认设置允许不经许可使用有版权的材料,这激起了对道德和法律风险的批评。 Despite safeguards, the app’s default settings permit copyrighted material use without permission, fueling criticism over ethical and legal risks. 公众日益怀疑AI, 调查显示对AI的伤害日益恐惧, Public skepticism toward AI is growing, with surveys showing rising fears about harm from AI, and backlash over low-quality, impersonal content and deepfakes.