Two years ago, Lorraine He (now a 24-year old law student) was told not to use AI in her assignments. To get around a ban on ChatGPT at the time, students were required to purchase a mirror site version from a secondhand market. It was widely used, but was often frowned upon. Her professors no long warn her students against using AI. They’re encouraged to use AI, as long as they follow best practice. She is not alone. Chinese universities are experiencing a quiet revolution, just like those in the West. According to a survey conducted by the Mycos Institute in China, a higher-education research group that specializes in higher education, the use generative AI is now almost universal on campus. In the same survey, only 1% of Chinese university students and faculty reported that they had never used AI tools for their studies or jobs. Nearly 60% of respondents said they use them frequently, either multiple times a daily or several times per week.
But there is a key difference. While many Western educators see AI as something they must managemore Chinese classrooms treat it as a skill that needs to be mastered. As the Chinese-developed DeepSeek model gains popularity worldwide, people increasingly view it as a source for national pride. In Chinese universities, the conversation has gradually shifted away from worrying about academic integrity and towards encouraging literacy, productivity, as well as staying ahead.
In public opinion, the cultural divide is more evident. Stanford University’s Institute for Human-Centered Artificial Intelligence, in a report examining global AI attitudes, found that China is the leader in terms of enthusiasm. Around 80% of Chinese respondents stated that they were “excited about” new AI services, compared to just 35% in the US or 38% in the UK.
Fang Kecheng is a professor of communications at the Chinese University of Hong Kong. She says that this attitude is not surprising. There’s a tradition in China of believing that technology is a key driver of national development, dating back to the 1980s when Deng Xiaoping said that science and technology were the primary productive forces.
Liu recommends students use generative AI for writing literature reviews, drafting abstracts, generating charts, and organizing thoughts. She has created slides that show examples of good and poor prompts along with a core principle: AI cannot replace human judgement. She says that only high-quality inputs and smart prompting will lead to good results.
Liu told her students that the ability to interact with machines was one of the most valuable skills in today’s society. “Instead of having students do this privately, we should discuss it in the open.”
The trend is growing across the country. MIT Technology Review examined the AI strategies of 46 of the top Chinese universities, and found that they had all added interdisciplinary AI classes to their general education curriculums, AI-related degree programs and AI literacy courses in the last year. Tsinghua is, for example establishing a new general education college that will train students in AI and another traditional discipline such as biology, healthcare or science.
Major universities like Remin, Nanjing and Fudan Universities are launching general-access AI degree programs and courses that are open to everyone, not just computer science majors, like the traditional machine learning classes. In 2024, Zhejiang University will require all undergraduates to take an introductory AI course.
Lin Shangxin told local media recently that AI is an “unprecedented chance” for humanities, social sciences and the arts. Lin told The Paper that AI, instead of being a challenge, would empower humanities study.
This collective action is a reflection of a central government initiative. In April 2025 the Ministry of Education will release new national guidelines that call for “AI+ education”a reform of all educational levels aimed at cultivating critical thought, digital fluency and real-world skill. Beijing’s municipal government mandated AI instruction in all schools, from universities to K-12.
Fang is of the opinion that formal AI literacy education can help bridge a growing divide between students. He says that there is a large gap in digital literacy. Some students are fluent with AI tools. Other students are lost.”
Creating the AI university
Due to the lack of Western tools such as ChatGPT and Claude on campus servers, many Chinese universities began deploying local versions DeepSeek in order to support their students. Many top universities have their own locally hosted version of Deepseek. These campus-specific AI system-often referred as the “full blood version” of Deepseek –offer longer contextual windows, unlimited dialog rounds, and broader functionality than free public-facing versions.
The trend is similar in the West where companies such as OpenAI and Anthropic offer campus-wide education tiers. For example, OpenAI recently offered ChatGPT Plus for free to all U.S. college students and Canadians, while Anthropic launched Claude for Education, with partners including Northeastern University and LSE. In China, however, the initiative is usually driven by universities rather than the companies themselves. Zhejiang University says that the goal is to give students full access AI tools to keep them up-to-date with the rapidly changing technology. Students can access the models using their ID.
Yanyan Li, Meifang Zhuo and other researchers from Warwick University, who studied the use of AI by students at UK universities, believe that AI literacy is crucial for students’ success.
Together with their colleague Gunisha AGGARWAL, they conducted a focus group including college students of different backgrounds and levels in order to learn how AI is used within academic studies. Students’ knowledge about AI is mainly based on personal exploration. “While students are aware that AI outputs are not always reliable, we saw a lot anxiety about how to use them correctly,” says Li.
Zhuo says that the goal should not be to prevent students from using AI, but rather to guide them in harnessing it for effective teaching and higher-order reasoning. This lesson has been slowly learned. A student from Central China Normal University, Wuhan, told MIT Technology Reviewhow, just a year before, many of his classmates had paid for mirror sites of ChatGPT using VPNs or semilegal online marketplaces in order to access Western models. “Now, everyone uses DeepSeek or Doubao,” said the student. “It is cheaper, it works with Chinese, and nobody worries about being flagged anymore.”
Yet, even with increased support from institutions, many students still feel anxious about using AI correctly or ethically. According to Rest of World’s report, the use of AI detection tools created a gray economy where students paid freelancers hundreds of yuan for them to “AI-detection proof” their writing. Three students told MIT Technology Review that this environment had created confusion, stress and increased anxiety. They all said that they appreciated it when their professor offered clear policies and practical advise, not just warnings.
The law student from Beijing recently joined a group to improve his AI skills in preparation for the job market. For many, learning how to use AI more effectively is not only a study hack but also a necessary skill on China’s fragile employment market. According to a report from the Chinese media outlet YiCai, AI-related skills will be a plus for 80 percent of job openings in 2025. Many students see AI as their lifeline in a slowing economy and competitive job market.
Zhuo says that “we need to rethink the definition of ‘original work’ at this age of AI”and universities are an important place for this conversation.
