What should universities teach when AI teaches?

As China's higher education system enters an era where artificial intelligence is not only a visitor in the classroom but also a co-teacher, universities face an important question: What should they teach in the age of AI?

At Tsinghua University, AI-assisted teaching has already taken root with the introduction of a human-machine collaborative learning model in the Department of Chemical Engineering. Other universities are not far behind. Northwest Agriculture and Forestry University uses AI to generate calculus images, Huaqiao University, capitalizing on the technology, builds basic coding frameworks for creating digital videos in traditional Chinese style and improving teaching and research, and Xi'an Jiaotong-Liverpool University has even experimented with cultivating AI-empowered translators.

What is striking is not the technology itself, but how it is being embraced. Teachers are not treating generative AI as a shortcut. Instead, they are wielding it as a tool for exploration. For instance, AI can draw up a plan for a low-carbon campus, complete with green walkways and energy-saving dormitories, within seconds. This is precisely where real education begins. Class time is saved for students to dig deeper by asking questions such as data sources, funding allocation and project execution strategies.

This transformation in China, where the vast higher education landscape encompasses both research universities focused on cutting-edge fields and vocational colleges serving local industries, is particularly relevant to universities overseas. In 2025, China explicitly called for integrating AI into teaching methods, textbooks and curricula, emphasizing the cultivation of independent thinking, problem-solving, communication and cooperation skills. This is not just about "initiating some AI-related courses" for students but rather about a fundamental question: What type of talent should universities develop?

Universities still need to impart knowledge, but not cramming. Instead, it's knowing what to trust. AI can quickly generate seemingly perfect explanations, but AI can hallucinate and present false information with convincing confidence. Without solid foundational knowledge in their majors, students may conveniently trust the false conclusions generated by AI. Therefore, education should focus on helping students understand concepts, trace sources, compare evidence and identify logical gaps.

This makes critical thinking more important than ever. For instance, AI can quickly design a comprehensive river management plan, but students must ask harder questions. How are upstream agricultural needs balanced with downstream urban demand, and how is harmony achieved between ecological protection and employment? While AI excels at optimization, education can help students develop skills to understand complex situations, gather evidence, draw sound conclusions, and identify flaws in arguments.

Notably, value judgments cannot be outsourced to machines. When resources are limited and interests conflict, the central issue is not speed but fairness. Who should bear the responsibility? If students only use tools without understanding real-world situations, ethical and institutional boundaries, as well as public responsibilities, their thinking may become more machine-like efficient but emotionally detached.

Cross-cultural communication skills are also becoming more important in the age of AI. The belief that accurate AI translation eliminates the need to learn foreign languages and cultures is a misconception. AI translation often misses cultural and social nuances. For Chinese graduates to effectively collaborate with European engineers, Southeast Asian businesses, African communities, and Latin American researchers on issues such as energy, agriculture and public health, they should not only translate but also clearly communicate Chinese experiences and understand others' perspectives.

Therefore, universities must transform from "knowledge factories" into "ability shapers". Four measures can help in this direction.

First, courses should be reorganized around real-world problems rather than merely labeled with "AI+". Law students can examine algorithmic responsibility; journalism students can scrutinize the trustfulness of AI-generated content; engineering students can use models to forecast risks while accounting for margin of error; and literature students can interrogate the limits of machine-authored texts. The real world does not organize itself by academic departments and neither should our curricula.

Second, the evaluation systems also need to change. If examinations only reward cramming, students will naturally seek AI assistance for writing. A better evaluation should focus on the learning process: how questions are asked, how AI tools are used and how students respond to different opinions in group discussions. Requiring students to document AI usage and learning logs may better safeguard academic integrity than simply relying on software to detect AI-generated content or plagiarism.

Third, the role of teachers must also evolve. AI cannot replace teachers' insight, understanding and ability to provide differentiated instruction for students. In the AI era, teachers are more like editors, coaches and ethical advisors helping students refine vague interests into researchable topics and give warmth and character to machine-generated drafts.

Finally, universities must remain open systems. Abilities are not developed in closed classrooms but are honed through projects, community service, corporate internships and international collaborations. In China, this means allowing students to use AI in specific scenarios such as urban renewal, rural vitalization and green energy, while also recognizing what AI cannot replace: trust, empathy, responsibility and long-term commitment.

Across East Asia, and especially in China, education has always been a national priority. The significance of AI-assisted teaching in Chinese universities lies not in technological novelty but in how a vast education system integrates innovation with equity, quality and development goals. The World Economic Forum pointed out in the "Future of Jobs Report 2025" that technological changes will profoundly reshape the global employment and skill structure by 2030. This means universities, whether in China, Europe, or Africa, must move beyond "how to teach faster" and confront the deeper question of "how to cultivate people to better discern, cooperate and take responsibility".

When AI takes on some teaching roles, universities will continue to conduct classes in mathematics, history, engineering, medicine, and languages. While AI provides answers, universities shape minds. Their classrooms must teach students to ask deeper questions, listen with empathy, articulate clearly and collaborate globally.

As AI advances, it becomes increasingly crucial for universities to cultivate skilled and responsible users of this powerful technology.

The author is dean of the School of Marxism, Southern University of Science and Technology, and former president of the University of International Business and Economics.

The views don't necessarily represent those of China Daily.

If you have a specific expertise, or would like to share your thought about our stories, then send us your writings at opinion@chinadaily.com.cn, and comment@chinadaily.com.cn.