The integrity of scientific publishing is being tested by the rise of AI-generated content. Researchers have used AI to create fake papers and submit them to IEEE, a leading scientific publisher, to see how well its human-review processes and research-integrity checker programs can detect them. The experiment, led by freelance writer and doctor of applied mathematics Benjamin Skuse, aimed to assess the effectiveness of IEEE's defenses against AI-generated papers. Skuse used AI tools to generate a paper that was more or less plausible and science-like, without making any significant changes to it. He then submitted the paper to IEEE, where it was reviewed by human gatekeepers and checked by research-integrity programs. The results showed that IEEE's defenses were able to detect the fake paper, but only after a thorough review process. The experiment highlights the ongoing challenge of verifying the accuracy of AI-generated content in academic research. As AI-generated content becomes more prevalent, it is essential to develop more effective methods for detecting and preventing the spread of fake or misleading information in scientific publishing. The experiment also underscores the importance of human review and verification in maintaining the integrity of the scientific record. While AI has the potential to revolutionize scientific research, it also poses new challenges for publishers and researchers. By testing the limits of AI-generated content, researchers can help develop more effective strategies for ensuring the accuracy and reliability of scientific publishing.