Core Viewpoint - The increasing capability of AI-generated content has led universities to implement strict measures for detecting AI usage in student theses, raising concerns about the reliability of such detection systems and the potential for misjudgment in evaluating original work [1][5]. Group 1: AI Detection Issues - Some universities require that students' theses have an AI detection rate below 15%, but there are reports of original work being flagged as having a high AI rate due to inconsistencies in detection results [1][4]. - A case was reported where a student's thesis showed a significant fluctuation in AI detection rates within a day, from 10.37% to 27.54%, leading to additional costs for revisions [1][2]. - A professor experienced similar issues, where content developed over three years was incorrectly flagged as AI-generated, highlighting the inadequacy of current detection tools [3][4]. Group 2: Expert Opinions - Experts acknowledge that misjudgments in AI detection are common and that the technology cannot completely eliminate errors, as it may not be trained on specific contexts [5]. - The standards for determining whether a paper is AI-generated remain vague, making it challenging for detection systems to accurately assess originality [5]. - Recommendations suggest that universities should not enforce AI detection as a mandatory graduation requirement but rather focus on guiding students in the appropriate use of AI tools [5].

看到今年的毕业季，估计王勃会气得从地里爬出来，把开发AIGC检测模型的人电脑插头拔了。 事情是这样的，今年很多学校为了防止学生用AI写毕业论文，在论文送审时加了一项"AIGC检测"，意思就是检测你文章里多少内容是由AI生成的。 因为自己写得太专业，被检测成AI写的。 有给导师改一遍，AI生成疑似度反而涨了的。 甚至有人把《滕王阁序》塞进去，发现其AI生成疑似度高达惊人的99.2%。 起初，这条公告并没有掀起什么波澜。 但直到最近，大伙的论文都接近尾声了，才发现这玩意儿让大伙的毕业季变成了走马灯。 "我写的内容会被识别成AI，而AI写的反而不会被识别成AI，因此AI检测论文就是在检测谁能把话说的更不像人话。"25届毕业生小蛋如是说，此处内含 脏话过滤器。 大伙发在网上的评论，更是一个比一个离谱。 这下合理了，原来咱勃哥开的是AI挂，我就知道没人能即兴写出这么好的文章。 而且你猜怎么着？AI生成疑似度的检测，也是由AI进行的。 这活儿，别说碳基生物了，硅基生物都想不到，自己被设计出来，要帮人写文章，改文章，还要帮你猜猜这文章是不是人写的？？早知道上辈子烂在服务 器里了。 为了和广大毕业生感同身受，咱也掏出了同事珍 ...