On May 28, Professor John E. Hopcroft and Professor Chen Baoquan delivered a lecture titled “Education in the AI Age” at the Yenching Academy hosted by Dean Dong Qiang. John E. Hopcroft is an A.M. Turing Award Laureate, Professor Emeritus in Computer Science at Cornell University and Visiting Chair Professor, Director of Center on Frontiers of Computing Studies (CFCS) at Peking University. Chen Baoquan is Boya Distinguished Professor and Associate Dean of the School of Artificial Intelligence at Peking University. This was the final lecture for the Topics in China Studies Lecture Series for the 2024‒25 academic year.
Interview Notes
Prof Hopcroft began with talking about what AI is, stating that there is no “intelligence” in AI, which solely consists of pattern recognition and statistics compressing large amount of data. But AI has had tremendous impact in helping other disciplines like biology and medicine and so on. The professor referred to Les Valiant’s research, clarifying that the aspect of intelligence that distinguishes human beings from other species lies in the human’s capability to educate one another. The ability of education allows humans to develop complicated technologies, such as the one allowing people to travel to the moon and come back to the Earth safely. Prof Hopcroft pointed that AI is going to change human societies and many goods and services that one needs will become automated. He cited the fact that the Americans were involved in agriculture accounted for 95% of the total population 100 years ago and that the figure is only 5% at present. The professor highlighted that all nations have to deal with such a social transition.
Prof Hopcroft emphasised that all new technologies have advantages and disadvantages. He cited automobile as an example, which brings both efficiencies and risks. Concerning information technology, he gave a case to show how vulnerable our private data is in the information era, that if he looked at all the searches made by one IP address, he would know an enormous amount about the person on that IP, such as the gender, the person’s hobbies, and even his or her itinerary in details. Therefore, every nation needs to guarantee that the data of their citizens cannot go out of the country. Prof Hopcroft highlighted that nations must establish laws to protect citizens’ privacy and ensure traffic safety in the face of technological advancements.
Prof Hopcroft gives first priority to helping students foster their computer talents and in education at large. He deemed it most important to focus on what one truly enjoys, not a high-payment job, in one’s career path. The professor advised new education pattern in China to lessen pressures on students. Also, he recommended a shift in university evaluation system, putting the quality of teaching as the core metric to evaluate a university in China, to foster creative students.
Prof Chen looked at the development path of AI in China, stressing on the synergistic relationship between algorithm innovation and computational power growth. He cited Deepseek as an example, which demonstrates that in the research on large language models (LLMs), high-performance algorithms can effectively enhance computational efficiency. According to Prof Chen, China is well-positioned to develop unique strengths in the field of algorithms, but it also has to deal with the increasing demands of computational power, as AI keeps developing; therefore, efforts are required equally for algorithm innovation and computational power growth, to gain a competitive edge in the global AI arena.
The professor went on to discuss how AI is changing our employment landscape and social structure. He noted that AI functions more as an “intelligent assistant”, replacing human workers in rule-based tasks, while humans remain superior to AI in areas requiring creativity, experience, and emotional connections. AI will reshape workflows and free up human workers from routine tasks. In this sense, AI may appear to reduce interpersonal collaboration; but in fact, it will catalyze higher-quality interpersonal relations, empowering humans to focus on the high-level collaboration that demands more emotional intelligence and creative thinking. Such transformation requires concerted efforts of the entire society to leverage AI for optimizing the social structure.
At the end of his interview, Prof Chen addressed the issue of AI-empowered higher education. He pointed that while AI is developing fast, education must shift from knowledge transmission to nurturing competencies. While recognizing the importance of the delivery of facts, Prof Chen emphasized that education must shift the focus onto equipping students with abilities of how to learn and how to think in an era where AI serves as an accessible reservoir of innumerable facts. By citing the ancient Chinese saying about “giving one a fish” versus “teaching one how to fish”, the professor stressed the importance of teacher-student interaction, peer communication, and inquiry-based learning. He suggested that higher educational institutions harness AI to both preserve the quintessence of traditional Chinese education and cultivate students’ capabilities for critical thinking and innovation, enabling students to navigate the changing world.
Review of the Lecture
Drawing on the profound impacts of AI on the humanities and social sciences, Prof Chen Baoquan explored the opportunities and challenges posed by technological advancements. He invited Prof John Hopcroft to review the milestones in the history of computer science and AI development, to address the concerns of parents, students and education professionals.
Prof Hopcroft drew a concise timeline of AI evolution. Mathematicians and physiologists of the 1950s started to study “pattern recognition”, laying a theoretical foundation for AI. In the 1960s, Cornell University made a breakthrough in computer recognition, capable of classifying patterns, and laid theoretical foundations for the development of Support Vector Machine (SVM) as an early core instrument for AI. The Deep Learning Revolution started in 2012, enabling a sky-roaring development of computer vision and related sectors, facilitated by the availability of massive, labelled image datasets and the powerful parallel computing with GPUs. He noted that the Turing Award, often referred to as the “Nobel Prize of Computing”, has witnessed AI’s path from lab to real life; however, the professor reiterated that technological innovations must always serve humans’ needs.
Prof Chen traced the trajectory of technological evolution, “GPUs are the ‘unsung hero’ in the revolution, if we see the symbiotic history of AI and technology from the perspective of computing graphics.” The GPU was initially designed as an image rendering tool, the powerful paralleling computing of which enables the high-performance processing of matrix transformation and light/shade simulation in video games. The GPU developers soon were aware that in addition to image rendering in virtual environments, GPUs could accelerate scientific computing. This capability paved the way for GPU. CUDA paralleling computing architecture created by NVIDIA allowed GPUs to break through the boundaries of image processing and become the “computing engine” for machine learning models. Computing Graphics laid the fundamental bedrock for computational capacity for AI, which in turn catalyzed the intelligent image generation. For example, the real-time rendering of Digital Humans was computationally intensive. At present, GPU clusters and deep leaning algorithms enable virtual characters with both millimeter-level facial expressions and contextual reasoning. It is achieved by a synergistic co-evolution of GPU computational power and AI architectures.
Next, Prof Chen raised the question “Will AI become intelligent?” To this, Prof Hopcroft explained, “The only ‘intelligence’ related to AI is in the name. It consists solely of pattern recognition and statistics compressing of large amount of data.” The professor noted that the pattern recognition technology can help in many disciplines, such as medicine and agriculture and it has changed tremendously the landscape of computing science; yet, AI is not “being intelligent” as humans.
By invoking the lecture by Michael J. Sandel, Professor of Government at Harvard University and eminent political philosopher, delivered at Peking University, Prof Chen pushed forward a discussion of the challenge to “human nature” posed by AI. Prof Sandel showed the example of The Beatles’ AI recreation of John Lennon’s unfinished song, which aroused a heated discussion. Prof Chen then asked Prof Hopcroft, “Will AI surpass human intelligence someday?”
Prof Hopcroft gave his opinions from the perspective of cognitive science, that before we discuss what “intelligence” is, we must admit that researchers have no definition of “intelligence”. On the one hand, it’s hard to define “intelligence”. Humans tend to define intelligence with metrics such as logic and learning capability. An untrained dog gets frozen when confronted with an unseen danger in the forest, while humans may react promptly. However, the capability of dealing with the unknown may not be universally present in all human beings, let alone being replicated by AI. Also, error correction is another issue. After making errors in image recognition, humans are able to reflect on the errors and make adjustments, while AI’s error correction relies on labelled data and algorithm iteration and is in essence statistical optimization, without genuine understanding of the mistakes.
Prof Hopcroft continued the conversation discussing his own opinions on education, “The purpose of education is to find out what you enjoy and steer your career towards it.” As for the involution and consequent anxiety that prevail among Chinese students, Prof Hopcroft advised the students to stop for a while and ask themselves about what they truly enjoy. He said that many Nobel Laureates he knew have succeeded not because of the innate talent, but their decades-long dedication to what they enjoy.
Prof Hopcroft cited his own life experience as an example. He grew up in Seattle and never planned to be a computer scientist. He was only curious about mathematics and mechanisms. What he did was grasping every opportunity to try. And in opportunities he found out possibilities. He noted that life is not a scripted play and that one may miss crucial opportunities due to overplanning.
The next issue may be whether AI will fundamentally change education? For this, Prof Chen cited the experimenting AI-led instruction in a Kansas school, where students engage with two hours’ interaction with AI every day and participate in school sports, handicraft and other group activities or engage in independent exploration in the rest of the day. All the courses are entirely individually tailored, and the system tracks each student’s learning progress in real time. He then asked Prof Hopcroft for his opinion.
Prof Hopcroft found the experiment stripped education of its core, that is, education should be human-centered. Even though AI is perfectly capable of delivering knowledge to students, it cannot replace a teacher, because a teacher cares for the growth of students as “humans”. The professor noted that AI could merely adjust its algorithms and parameters to address students’ perplexion, while a teacher thinks about students’ changes. The emotional connection is the most precious essence of education.
Prof Hopcroft attributed the “change” of education to two tools—blackboards and printing. The blackboard functions as the platform where orally-delivered knowledge becomes visualized, allowing for group study. Printing has broken up knowledge monopoly, enabling the general public to access knowledge. The professor once thought that TVs would bring radical changes of classroom teaching; yet, he later found the TV was nothing more than a tool for one-way instruction. AI may appear to be a challenge to traditions, but in essence, it depends on image recognition to make judgment. The overemphasis on AI’s role and function represents a thinking of the primacy of instrumental rationality.
Prof Hopcroft also shared his thought on educational space, in many universities, teachers’ offices are located in the administration building, while classrooms are often in more peripheral areas. An invisible wall rises. If, teachers’ offices are situated on the third and fourth floors, while classrooms are on the first and second floors in the same building, then teachers would pass the classrooms every time they go to buy a cup of coffee. And teachers and students would meet and have a chat and even an impromptu discussion. Students can have more interactions with their teachers. And education become a natural process in ‘chance encounters.
In the Q&A Session, Prof Hopcroft gave his opinions on “whether a child should be allowed more e-devices”. He distinguished TVs from iPads in terms of impacts on children. Children passively receive one-way information from TV. But in the 1950s, TV helped unveil the world in front of the children in remote areas. Indiscriminate prohibition of TV watching only produces adverse effects. On the other hand, iPads have had profound impacts on children, particularly the under 3s. The 0-3 age group are in the process of establishing their cognitive capability for ‘how to learn’; therefore, premature exposure to e-devices can impair children’s cognitive development in the long term. In addition, Prof Hopcroft addressed other issues raised by our Scholars, from AI’s role in Chinese learning, “intelligence” and “empathy” of AI, to Prof Hopcroft’s “China story” of conducting research on AI in China.
In an era where AI can compose a song, drive a car, or even reconstruct the face of the deceased, this lecture warned us that the true danger may not be the human-like machine, but rather the machine-like humans. The more powerful the technology, the more conscious we must be of its limits. In an era, full of uncertainties, neither the blind faith in technology nor sticking to convention leads to the future of education. Perhaps, it’s more important “to become oneself” than “to become somebody”. The purpose of education in the AI age may help people to be firmer about the reason why we exist and why we are irreplaceable. The most important thing for education is to care for the fragile yet radiant part of humanity: the courage to face the unknown, the impulse for beauty, and the genuine understanding of mistakes.
