Unverified50% confidenceFactExact time
Deepfake technology first emerged around 2017, with early implementations based on Generative Adversarial Network (GAN) architectures.
1
Sources
50%
Confidence
Long-term
Relevance
7/24/2026
First Seen
Sources
一句提示词生成电影海报，AI图像one-shot能力解析
redditr/ChatGPT7/12/2026
Related Claims
VerifiedDeepfake technology can be traced back to 2017, when it used GANs (Generative Adversarial Networks) to achieve face swapping.82% similarUnverified深度伪造技术的技术底座经历了从生成对抗网络（GAN）到扩散模型（Diffusion Model）的代际跃迁80% similarUnverifiedDeepfake technology is based on generative adversarial networks (GANs) in deep learning, and its production cost has dropped from hundreds of thousands of dollars several years ago to a level affordable by ordinary individuals.72% similarUnverifiedThe core principle of deepfakes relies on Generative Adversarial Networks (GANs), in which a generator and a discriminator compete against each other to continuously optimize output quality.69% similarUnverified深度伪造的核心原理基于生成对抗网络（GAN）或扩散模型，早期需要数百张照片和数小时训练时间69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/608953API
curl https://kongchang.com/api/v1/knowledge/claims/608953MCP
get_claim(id=608953)