Recently, the 6th Seminar of the Ministry of Education's Online Education Research Center and the 10th Anniversary Conference of the Ministry of Education's Online Education Research Center, hosted by Tsinghua University, was held at Tsinghua University. Professor Luo Sanzhong from the Department of Chemistry at the university was to participate in a roundtable discussion. However, before that, he needed to finish teaching his undergraduate class for the day.
During the class, Professor Luo posed a question to the students—could they create a machine that could learn organic chemistry together with them?
In the fall semester of 2023, Tsinghua University launched a teaching reform plan that integrates artificial intelligence (AI) into education. The plan, starting with 8 courses as pilot programs, developed AI teaching assistant systems to provide students with 24-hour personalized learning support, intelligent assessment, and feedback.
Building on this initiative, the university is set to offer 100 courses integrating AI into teaching, utilizing AI to assist or deeply engage in the courses, creating AI teaching assistants and AI teachers to enhance classroom teaching efficiency and quality.
The "machine" mentioned by Professor Luo is this kind of artificial intelligence.
After finishing his class, Professor Luo joined the forum where he discussed alongside three other Tsinghua teachers who were also experimenting with using AI in their teaching. Their topic of discussion was how AI can empower curriculum development.
Encouraging students to see AI as a "learning companion"
Professor Luo has been asking students to design AI programs since 2021. Several students approached him with enthusiasm to try it out.
"Now, those students have become core members of the research team," Professor Luo said. Over the past two years, he has made the same request to each new group of students. This request has now evolved—can students teach AI about chemistry?
"We have made similar attempts, with some students trying to teach AI about organic chemistry," Professor Luo said. Teaching AI requires students to have a complete mastery of the knowledge, enhancing their proactive learning. "This is a typical application of AI in promoting education."
Professor Lu Diannan from the Department of Chemical Engineering at Tsinghua University also initially hoped that AI could learn alongside students.
His course was included in the first batch of 8 courses using AI-assisted teaching as internal pilot programs. When discussing this with students, he encouraged them not to see AI as a "teacher" or an "assistant" but as a "learning companion." "In other words, AI is just a learning partner in the class, observing whether I teach it well or if real students learn better from me."
This was Lu Diannan's initial intention in involving AI in teaching. However, after over half a year of practice, he admitted that his perspective had significantly changed.
"After extensive vertical training, we found that current AI models can deduce formulas, expand knowledge, and even help teachers fill knowledge gaps," he said. He now believes that generative AI is sufficient to enhance students' learning, productivity, and innovation within his field of expertise.
Lu Diannan stated that in the current context, teaching knowledge is no longer the most crucial task for fostering innovative talents; the primary task is to help students transform knowledge into skills. "Once they acquire these skills, they need to demonstrate them by completing specific tasks. If their knowledge is incomplete, AI intervention can help students fill those gaps and assist them in completing the tasks."
In response, Li Ning, a distinguished professor at Tsinghua University's School of Economics and Management, emphasized that the integration of AI with traditional classroom teaching comes in various forms, but AI should not be simplistically viewed as a mere "auxiliary tool." "We need a comprehensive curriculum design that allows humans and AI to become close collaborators. This is a key factor in determining whether AI can truly integrate into education."
Teaching in the era of artificial intelligence
During the discussion, almost all teachers acknowledged that with the introduction of AI into the classroom, the boundaries between teachers and students are no longer as clear-cut as before.
"In the traditional teacher-student binary scenario, the teacher-student relationship easily falls into a pattern of 'teachers teach, students learn,' where no matter how much you encourage, students instinctively rely on the teacher's authority in learning," said Xu Luping, Vice Dean of Tsinghua University's Xingjian College.
However, with the introduction of AI, the classroom becomes "interesting"—it alters the original ecosystem of the classroom, turning it into a scenario of triadic interaction among students, teachers, and AI.
"In this scenario, teachers are no longer absolute authorities, as the knowledge acquisition capability of AI may even surpass theirs. As a result, the teacher-student relationship shifts from a simple 'teaching' and 'learning' relationship to a peer-like relationship of mutual exploration and discussion. This also allows students to naturally enter a state of flow where they are more willing to ask questions and put forward ideas," Xu Luping explained.
For teachers, this change undoubtedly poses a challenge.
Li Ning noted that during class, he found that teaching certain knowledge points seemed to have little significance. "Because AI may explain things better than us, and students can delve deeper by continuously asking questions." He raised the question: How should we teach students when we assume that all students will learn using AI? This is a question that all teachers need to contemplate. In addition, during a study, Li Ning discovered a phenomenon - the era of artificial intelligence has a reshaping effect on human capabilities.
"Simply put, the students we educate will go on to work in various fields. However, individuals who perform well and excel in professions in a non-artificial intelligence environment may not necessarily perform better in an artificial intelligence environment; and vice versa," he explained.
This phenomenon will require universities to consider what abilities to impart to students in order to enable them to have better career development in new environments. "Currently, there is no clear answer to this question," Li Ning said.
Apart from changes in the role of teachers themselves, the state of students is also evolving.
"Our biggest confusion and challenge may lie here," Luo Sanzhong stated. If around the year 2000, the teaching targets of university teachers shifted from offline students to "netizens," then in the next three to five years, university teachers will face educational issues with "artificial intelligence natives." "This is also something we haven't thought through properly."
Education cannot be dragged along by technology
These unresolved issues require university teachers to explore them diligently in practice.
Luo Sanzhong mentioned that when dealing with issues related to artificial intelligence, teachers must not observe from afar but must observe closely to understand what problems it is actually solving, interact with it, and gradually deepen their understanding of artificial intelligence in the process.
In the classroom, there have been students suggesting to Luo Sanzhong multiple times whether a course on "Artificial Intelligence + Chemistry" could be offered. To this, his response was, "I first need to study it well."
"Compared to traditional teaching methods, a key feature of teaching in an artificial intelligence environment is to transform all passive learning into active learning, or in other words, to create a certain technological atmosphere that allows students and even teachers to involuntarily enter a state of learning, which places much higher demands on teachers," Luo Sanzhong said.
This viewpoint resonated with many teachers. For example, Li Ning bluntly stated that teachers should be the most proficient in using artificial intelligence in the classroom. Only then can he lead students to make various attempts. "On this issue, we cannot rely on students' creativity."
Of course, proactive attempts by teachers also require some hardware conditions and platforms, which became the second topic that sparked unanimous agreement.
"Technical support is crucial," Luo Sanzhong stated, emphasizing that regardless of the type of artificial intelligence large model, there will be demands in terms of computing power. In this scenario, if everyone can provide mutual support for artificial intelligence through a certain platform, it can optimize the allocation of educational resources and funds, further promoting fairness in education and teaching.
Xu Luping also believes that in the current rapid development of artificial intelligence technology, how to truly empower education with technology, rather than dragging education along, requires deep interaction and cooperation among technology experts, educators, and educational researchers.
"On the issue of empowering teaching with artificial intelligence, whether it is the technology itself or the cultivation philosophy and methods, it is global. Therefore, we should establish broader, more comprehensive, and deeper cooperation on a new starting point and platform globally to promote large-scale innovative talent development," Xu Luping said.
