Gu Jinjin, Wang Xinlei, Li Chenang, Zhao Junhua, Fu Weijin, Liang Gaoqi, Qiu Jing
School of Electrical and Information Engineering, University of Sydney, Sydney, NSW, Australia.
School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen, China.
Patterns (N Y). 2022 Jul 8;3(7):100511. doi: 10.1016/j.patter.2022.100511.
Destroying image integrity in scientific papers may result in serious consequences. Inappropriate duplication and fabrication of images are two common misconducts in this aspect. The rapid development of artificial-intelligence technology has brought to us promising image-generation models that can produce realistic fake images. Here, we show that such advanced generative models threaten the publishing system in academia as they may be used to generate fake scientific images that cannot be effectively identified. We demonstrate the disturbing potential of these generative models in synthesizing fake images, plagiarizing existing images, and deliberately modifying images. It is very difficult to identify images generated by these models by visual inspection, image-forensic tools, and detection tools due to the unique paradigm of the generative models for processing images. This perspective reveals vast risks and arouses the vigilance of the scientific community on fake scientific images generated by artificial intelligence (AI) models.
破坏科学论文中的图像完整性可能会导致严重后果。图像的不当复制和伪造是这方面两种常见的不当行为。人工智能技术的快速发展给我们带来了有前景的图像生成模型，这些模型可以生成逼真的虚假图像。在此，我们表明，此类先进的生成模型对学术出版系统构成威胁，因为它们可能被用于生成无法有效识别的虚假科学图像。我们展示了这些生成模型在合成虚假图像、剽窃现有图像以及故意修改图像方面令人不安的潜力。由于这些模型处理图像的独特范式，通过目视检查、图像取证工具和检测工具来识别由它们生成的图像非常困难。这一观点揭示了巨大的风险，并引起了科学界对人工智能（AI）模型生成的虚假科学图像的警惕。
