With the iterative development of artificial intelligence technology, more and more college students are starting to try using AI to search for information, polish sentences, and even use AI to write papers. In the first half of this year, universities such as North China Electric Power University, Hubei University, Fuzhou University, Nanjing University of Technology, Tianjin University of Science and Technology, and Communication University of China successively issued notices, stating that they will try to include detection of the risk of using generative AI in their undergraduate thesis review process. As one of the first universities in China to attempt to use AI to detect AI, North China Electric Power University has collaborated with CNKI to develop the "AIGC (Generative Artificial Intelligence) Detection Service System", specifically designed to detect the use of AI in the thesis of its 2024 graduate students. Zhang Lei, Vice Dean of the Graduate School of North China Electric Power University, told China Youth Daily and China Youth Network reporters that the detection rate of part-time graduate students is also high, with some papers reaching as high as 40%; In the classification of disciplines, the detection rates of economics, foreign languages, and humanities are relatively high, with some papers reaching 20% -30%. In contrast, there are very few engineering disciplines with detection results exceeding 10%. However, at present, North China Electric Power University has not yet provided a clear definition of how much the detection rate will not be defended. Zhang Lei told a reporter from China Youth Daily and China Youth Network that the main reason is that the current AI testing tools cannot guarantee the accuracy of the test results 100%. Therefore, the school has not yet introduced mandatory regulations, but instead provides the test results to the supervisor and defense committee as a reference. Shang Junjie, Executive Director of the Learning Science Laboratory at the School of Education, Peking University, believes that the transformation of education through generative artificial intelligence is an inevitable trend. The development of artificial intelligence technology can to some extent force education reform, and it should not be seen as a flood beast. More importantly, it is important to explore a path for the integration of new technology and education as soon as possible. From manual ghostwriting to artificial intelligence ghostwriting, "In the past, academic misconduct in academic paper ghostwriting was mostly done manually, but now artificial intelligence ghostwriting has emerged, which is also a clear indication of academic misconduct." Zhang Lei said. The principle of generating papers using artificial intelligence relies on deep learning, natural language processing, and generative models. It does not think word by word or search for sources of information item by item like people do when writing, but generates texts through computation and prediction. During the generation process, the model also takes into account the previously generated content to ensure the coherence and consistency of the entire text. Compared to manual ghostwriting, artificial intelligence ghostwriting has absolute advantages in efficiency and convenience. Yang Yu (pseudonym), a student majoring in foreign languages at a university in Beijing, tried using artificial intelligence to generate his graduation thesis in March this year. In less than 10 minutes, he completed an undergraduate thesis of about 8000 words. "This is a level that I can't even rack my brains to write." Yang Yu said, but upon careful examination, he found that names such as "Zhang San" and "Li Si" appeared in the references. Yang Yu was helpless and decided to give up this shortcut. After more than two months, he followed the framework provided by AI, independently researched materials and literature, supplemented content, and completed his graduation thesis again. At the end of 2023, China Youth Daily and China Youth School Media launched a questionnaire survey on the use of AI tools among university students nationwide, collecting a total of 7055 valid questionnaires. The survey results show that 84.88% of respondents have used AI tools. AI tools have been applied by respondents in various learning and life scenarios, such as information retrieval (61.30%), translation (58.31%), and writing (45.75%). 77.51% of respondents believe that AI tools can improve work and learning efficiency to a certain extent. "Using AIGC technology to assist in thesis writing may lead to students developing an 'artificial intelligence dependency' that weakens learning autonomy and to varying degrees affects the fairness of talent evaluation." Associate Professor Li Zhikai from the Law School of Guangxi Normal University believes that there may even be risks of suspected plagiarism and plagiarism. At the same time, most students find it difficult to determine whether the content generated by AI comes from materials protected by intellectual property rights or is mixed with false information. This makes it easy for students to include false information in their thesis if they are not careful when using AI to assist or even write papers on behalf of others. Li Zhikai told a reporter from China Youth Daily and China Youth Network that if a student cites AI generated text that infringes on the intellectual property rights of others in their paper, they themselves will also face civil risks of infringement, and may even be recognized by the school as falsifying their paper, which may affect their degree acquisition. Research has found that under the influence of training data, the output content generated by AIGC is prone to stereotyped responses reflecting cultural and political biases. Returning to the process of writing a thesis, Li Zhikai believes that AIGC may provide some biased content as answers to students and affect their emotional cognition, attitude, and behavior. At the end of 2022, after the release of ChatGPT, Zhang Lei discovered that students began using AI to process text and images. "It is indeed very efficient, so we did not object at first," he said. Later on, more and more students used AI to generate homework, articles, and algorithms, and "we felt like there was a problem," Zhang Lei said. "The boundary between AI assistance and AI ghostwriting is not to eliminate new technologies, but to guide students to use them correctly," said Zhang Lei. For Sun Hao (pseudonym), who is studying Data Science and Big Data Technology at a 985 university in Beijing, AI is his auxiliary tool. When completing the experiment and submitting the experimental report, Sun Hao will hand over some of the work to AI. For example, in the task of translating code in the experimental steps, he will first send the code he has written to ChatGPT, let ChatGPT read the code he has written and write the process of thinking, and then correct errors himself, "saving me time typing word by word," said Sun Hao. In terms of experimental results, it is necessary to conduct some descriptive statistical analysis on the data, which Sun Hao considers to be his weakness. He will first write a paragraph, along with the data results and background, and send it to ChatGPT to generate a result analysis of the data. Then, he will combine his own writing with the AI generated part and finally put it in the experimental report. Sun Hao believes that AI only helped him enrich the content of the experimental report, expand and polish it, but fundamentally, the experimental content and findings were still completed by himself. Qiu Yinchen encourages students to use AI reasonably in their paper writing. As an associate professor at the European College of Beijing Second Foreign Languages University and a member of the artificial intelligence and language cognition laboratory, Qiu Yinchen has been exploring the application and implementation of artificial intelligence in the field of language translation for nearly 10 years. He found that compared to search engines, the new generation of conversational artificial intelligence tools have significant advantages in translation, data collection, listing cutting-edge research progress, and extracting central ideas from literature. Recently, at the 9th Party Congress of the Beijing Second Foreign Language Institute, where Qiu Yinchen is located, he proposed to "empower foreign language teaching with artificial intelligence.". This school is trying to integrate some basic courses with AI, introducing AI models and tools such as deep learning, voice evaluation, automatic correction, and intelligent language partners. It is trying to use AI to help students correct their spoken language, identify basic errors in homework, answer questions about subject knowledge and teaching management, and even provide students with an AI learning assistant to recommend personalized learning tasks and dynamic learning plans. "AI can provide a more personalized and humanized learning environment," said Qiu Yinchen. Last August, in the draft of the Academic Degrees Law of the People's Republic of China, "artificial intelligence ghostwriting" first appeared and was included in the prohibited category of academic misconduct. But on April 26th of this year, the 9th meeting of the 14th National People's Congress Standing Committee voted to pass the "Academic Degrees Law of the People's Republic of China", deleting "artificial intelligence ghostwriting" and retaining the word "ghostwriting". Li Zhikai told a reporter from China Youth Daily and China Youth Network that this indicates that it is currently difficult to simply determine whether the use of AIGC technology complies with the legal norms for academic dissertations, and existing laws do not make prohibitive provisions on assisted writing techniques. For academic dissertations, while meeting the legal norms for academic dissertations, AIGC technology can be used to assist in academic dissertations, but cannot be used for ghostwriting. "The advantages and risks of AIGC technology coexist." Li Zhikai said that abandoning food due to choking cannot prevent the popularization of AIGC technology. Even in the draft of the Academic Rank Law of the People's Republic of China, it only prohibits "ghostwriting" rather than "application". The most important thing at present is to explain and clarify the boundaries of academic misconduct in "AI ghostwriting", and establish scientific norms and guidelines to ensure the effective use of students. "The conflict and integration of AI and higher education" In the field of education, the transformation of social structure and talent demand brought about by technology will directly lead to the reshaping of the education system: first affecting the training objectives of the education system, and then affecting the training methods of teaching, learning, and teaching management. "Shang Junjie said that specifically in the technology integration represented by AIGC, repetitive knowledge production work will be replaced, and in the future, talents need to highlight the uniqueness of human beings more. In April this year, the Academic Affairs Office of Tianjin University of Science and Technology issued a notice stating that if the results of generative AI testing in undergraduate thesis exceed 40%, the school will issue a warning to students and require them to self check and correct. That month, the Undergraduate School of Communication University of China issued a notice on strengthening the standardized use of artificial intelligence management in the 2024 graduation thesis (design). Subsequently, multiple colleges of the school also issued relevant regulations and rules. Among them, the Continuing Education College of the school requires students to clearly disclose whether they are using AIGC in the detailed rules. If used, please provide detailed instructions on how to use it, including the model/software/tool name, version, and usage time. The auxiliary generated content involving factual and viewpoint citations must clearly explain its generation process and be specifically marked in the corresponding position of the graduation thesis (design) to ensure authenticity, accuracy, and respect for the intellectual property rights of others. Zhang Lei told China Youth Daily and China Youth Network reporters that North China Electric Power University will explore and research more specific management methods similar to plagiarism checking in the second half of this year, clarifying the proportion of AI generated detection rates that need to be modified or not defended. Meanwhile, different disciplines are also different, and different management methods should be developed. "But the commonality is that the core ideas, viewpoints, and experimental data must be proposed by the students themselves," said Zhang Lei. As a teacher, Qiu Yinchen believes that we should be good guides for students and academic gatekeepers - teaching them how to do real research, write original papers, and uphold the bottom line of academic ethics. In the foundation class of freshman year, Qiu Yinchen will incorporate research topics, preliminary ideas, and research methods into the explanation of knowledge points, and invite interested students to join the research on this topic in the classroom. In his opinion, early involvement of students in the research process can effectively cultivate their critical thinking and independent thinking abilities. With the ability of critical thinking and independent thinking, when producing academic papers, there is no need to worry about students having empty minds and no way to start. In Qiu Yinchen's view, as long as the innovation and logic of the paper are grasped, the detection of AI ghostwriting papers can be achieved. Shang Junjie believes that teachers not only need to bear the dual responsibility of teaching and educating students, but also need to play a leading role in using AIGC to assist students in learning, and strictly regulate the application of AIGC,