深度伪造Deepfake
Deepfake
用生成模型造出以假乱真的真人音视频：眼见不再为实。Generative models producing indistinguishable fake video and audio of real people: seeing is no longer believing.
深度伪造指用深度学习合成或篡改的图像、音频、视频。人难以辨真伪，且常以真实人物为对象。
综述指出：deepfake 算法能造出与真实影像无法区分的假内容。
检测与生成是一场持续军备竞赛，检测器普遍滞后于生成器。
滥用场景包括冒充真人诈骗、伪造事件影像。
Deepfakes are deep-learning-synthesized or manipulated images, audio, or video that humans struggle to distinguish from authentic ones, often targeting real people.
Surveys: deepfake algorithms can create fake media indistinguishable from authentic media.
Detection and generation are an ongoing arms race; detectors generally lag generators.
Abuse scenarios include impersonation scams and fabricated footage of events.
看到「实拍」视频先问一句有没有可能是合成的。涉及钱和决策的音视频，要换渠道二次确认。When you see 'real footage', ask first whether it could be synthetic. For audio and video involving money or decisions, verify through a second channel.
去导航：找现成的工具、平台、评测与学习资源Directory: ready-made tools, platforms, benchmarks and courses ↗
回到名词地图看第2站的其他 35 个词 →Back to the map: the other 35 terms in station 2 →
- Vasco 等《Deep Learning for Deepfakes Creation and Detection: A Survey》Survey: Deep Learning for Deepfakes Creation and Detection ↗