“高度逼真”
用英语怎么说？

OpenAI’s Sora application briefly offered users a novel capability: generating convincing self-deepfakes. By simply uploading a brief facial recording, users could insert their digital replicas into any conceivable scenario. Although the app was shut down after just six months, its brief existence exposed a profound vulnerability in human cognition. Some heavy users began experiencing a bizarre phenomenon where they started remembering events that had never actually occurred, but which they had watched themselves perform in AI-generated videos. For instance, after repeatedly viewing a synthetic video of himself scaling Mount Rushmore, a user named Andrew Deutsch felt a distinct flicker of recollection-a subtle neurological glitch-suggesting he had actually climbed it. This warning highlights that generative media can interfere with our autobiographical memory, which forms the core of personal identity, because our brains are not naturally equipped to reject highly realistic, self-referential mental imagery.

OpenAI’的Sora应用程序曾短暂地为用户提供了一种新颖的功能：生成令人信服的自我深度伪造视频。通过简单地上传一段简短的面部录像，用户就可以将自己的数字复制品插入到任何可以想象的场景中。尽管该应用在短短六个月后就被关闭了，但它的短暂存在暴露了人类认知的一个深刻脆弱性。一些重度用户开始经历一种奇怪的现象，他们开始记住那些从未真正发生过的事情，但他们曾在人工智能生成的视频中看到自己做过这些事。例如，在反复观看自己攀登拉什莫尔山的合成视频后，一位名叫安德鲁·多伊奇的用户感到了一丝清晰的回忆--一种微妙的神经系统暗示--这表明他实际上攀登过。这一警告强调，生成式媒体会干扰我们的自传体记忆，而自传体记忆构成了个人身份的核心，因为我们的大脑没有天生拥有抵御高度逼真且涉及自我的心理图像。