Nanjing University on Wednesday denied online claims that its smart classroom system monitors students' "head-up rate" or collects facial biometric information, after screenshots of the system sparked debate over privacy and algorithmic surveillance.
"Head-up rate" refers to the proportion of students looking at the teacher or blackboard rather than down at their phones or desks. Some universities have used the metric to assess classroom engagement.
The controversy began after students received notices that the university would use its smart classroom system, upgraded over the summer, to record attendance in certain courses for a week.
A screenshot subsequently circulated online showing surveillance footage alongside statistics including the number of late arrivals, attendance rate and head-up rate. Some teachers reportedly told class groups that attendance data would directly affect students' regular grades, further fueling the controversy.
Media reports said the project involved an investment of more than 31.52 million yuan ($4.7 million) and the upgrading of 113 public classrooms. Bidding documents listed functions including facial detection, facial recognition and AI-based head-up rate calculations.
Students said they had not been consulted in advance and expressed concerns that their facial information could be accessed by the university and technology companies. Even without precise facial recognition, they said, monitoring head-up rates alone made them uncomfortable.
In a statement posted on its official WeChat account on Wednesday, the university's undergraduate school and information construction management center addressed several claims, including that attendance rates would be linked to regular grades and that the midterm teaching inspection had been ordered by higher authorities.
The university said the renovation was intended to improve classroom audio and video facilities, enable livestreaming of courses and cross-campus interaction, and support course playback and the reuse of teaching data. It added that the project had followed government procurement procedures.
The midterm teaching inspection was routine administrative work, the university said, adding that claims it had been ordered by higher authorities and that attendance would be linked to regular grades "are not true".
Addressing privacy concerns, the university said: "Without authorization, our smart classroom system does not collect biometric information such as facial data or conduct facial recognition in teaching applications." It added that the system no longer collects head-up rate data.
The university pledged to respect and protect students' privacy while using digital technologies to improve learning.
The controversy has also prompted broader concerns among education experts about the educational value and privacy implications of AI-powered classroom systems that monitor students' head-up rates and collect facial data.
Gao Hang, vice-dean of the School of Education at Renmin University of China, said using head-up rates to judge students' engagement was "one-sided" and lacked academic support. Looking down, for instance, does not necessarily mean a student is not learning, as they may be consulting materials or taking notes, he said.
"Does a student have to stare at the teacher to be learning?" Gao asked.
He warned that relying on such indicators could lead universities to prioritize formalistic measures over assessing what students actually learn and internalize.
AI can improve education in many ways, including boosting classroom interaction, enabling real-time quizzes and polls, and recommending reading materials. However, in this case, the technology is being used mainly for management and surveillance, Gao said.
Li, a graduate student in Beijing, also criticized the system as "formalistic". She expressed concern that large screens could capture not only students' faces but also their phone screens and other personal information.
"We were never asked for consent, yet our privacy is fully exposed," she said, arguing that the system creates an appearance of effective management without genuinely improving teaching or learning quality.
"At university, if the content is truly interesting or useful, students will naturally look up," she added.
Yu Xinyan contributed to this story.
