“ID: 000214, State 1: Listening”
“ID: 000015, State 4: Writing”
“ID: 002045, State 5: Sleeping”
On the teacher’s screen is a live video feed of her classroom. On the monitor, a colored box is affixed to each of the student’s faces, tagged with an ID number and a state of attention based on an algorithmic interpretation of the student’s behavior. This information is then stored and analyzed to produce a weekly report posted on a mobile app for parents, school administrators, and teachers to view.
Imagine a world where every second of every day, a student’s level of engagement and emotion would be monitored. Above the blackboard, in front of each classroom and throughout the school, hundreds of surveillance cameras are installed to interpret students’ facial expressions as anger, fear, disgust, surprise, happiness, sadness, or neutral. The technology can even detect body movements to determine whether they were paying attention or not. Elementary school children wear headbands to monitor their level of focus and the data is sent in real-time to teachers so that they can ensure their pupils are paying attention and enjoying the lessons in class. Although all this may seem like a plot right out of a sci-fi movie, it is currently a reality in China.
In July 2017, the main governing body of China announced its goal of becoming the world’s leading artificial intelligence power by 2030. In the Next Generation Artificial Intelligence Development Plan (NGAIDP), China outlined its objective of building a domestic $150 billion AI industry. Consequently, schools, from elementary to universities, began to implement various types of AI-based technologies and softwares into their educational systems. At the Hangzhou No. 11 Middle School in eastern China’s Zhejiang Province, a “five-star smart campus initiative,” which involved facial recognition technology, was implemented into multiple aspects of school. Students could purchase lunch from the cafeteria and borrow books from the library without having to carry around ID cards. The technology removes the need for certain human administrative tasks such as entrance security and taking attendance. As part of the same AI initiative, schools began to create “smart classrooms” which include surveillance cameras that collect in-class behavioral data. The devices take a photograph every second and send the data to a computer server which uses an algorithm to sort the faces into one of seven emotions. The teachers can log in to the interface whenever needed to access the results. Red and green arrows point up and down to indicate when a student’s attention score rises or drops. A bar graph also shows how many minutes each pupil spends conversing, sleeping, or concentrating in class. Each class’ results are displayed on a large screen in the hallway for the entire school to compare. At Jinhua Xiaoshun Primary School in China, students wear headbands, using electroencephalography (EEG) technology, to monitor their focus and engagement level. A bright orange light on the wearable gear signifies a high level of focus. The information is collected and is organized into a graph that is sent directly to the teacher and posted in a chatroom for parents. Some parents and teachers have claimed that the class behavior improved and the technology is effective. This means that kids are being turned into straight-A students with an academically-focused mindset from a young age, or some might say, straight-A robots rather than students.
Although this technology may be useful for China in many ways, the system has its flaws. Dr. Theodore Zanto, director of the Neuroscience Department of the University of California, says that the technology “is prone to inaccuracies.” When the device is not set properly on the head, the detection of electrodes will vary. He had also stated that the devices could cause irritability and itchiness. Children at the Primary School expressed their discomfort in wearing the headbands. One student complained that the device squeezed on his head, leaving marks. In addition, although the technology might not work as intended, the pupils feel more pressured to pay attention in class. According to the Chicago Tribune, “Most students came to hate the constant monitoring — and the consequences that followed when the machines reported scores suggesting individuals or entire classes weren’t paying attention.” Pupils were not being motivated to learn out of their own desire to succeed but were only learning out of their fear of being punished by parents or teachers for poor grades. One student admitted that after the headbands were put in place, he never scored anything below a 95% on tests.
Only recognizing seven emotions, the facial recognition technology being used at Chinese schools could be flawed in equalizing body movement to attention. For example, a student who is looking down at his desk to find a pen could be marked as sleeping. Chen Hong, a Beijing resident who installs high-speed surveillance cameras for a living, was worried about the accuracy of the technology. The Chicago Tribune even stated, “Some students went so far as to figure out how to game the system by feigning what the cameras’ designers wanted to see.” One high schooler, Zhu Juntao, used his two forefingers to press up the sides of his mouth so he would appear to be smiling and he was never called out by the teacher for being moody.
Another concern for many parents was how the data was shared. Companies claimed that the data could go to government-funded research projects. Zhang Jiaojiao, one of the parents interviewed by the Wall Street Journal, says, “It’s for our country’s research and development. I don’t think it’s a problem.”
But not all would agree. “When we were young, we don’t want our parents to monitor us every day so if one person is being watched by an electronic eye every day, it must have a big impact on his or her psychological health,” Tian Guanghui, a concerned parent, said. Many others share Tian’s displeasure and his letter of complaint has gained over one hundred signatures. The Chicago Tribune reports, “…enough of them complained about what they saw as an intrusion on privacy that administrators last month hit the pause button on the cameras.” However, even after complaints were filed, China continues to insist on using AI in the classroom. Vice Principal Zhang Guanchao hopes to have the system back on campus.
Although China’s policy to incorporate AI in the classroom began in 2017, the Wall Street Journal’s article, “China’s Efforts to Lead the Way in AI Start in Its Classrooms,” in October 2019 brought more awareness to the topic. Other nations are chiming in to express their worries. Clare Garvie, an associate at the Center on Privacy and Technology at Georgetown University Law Center, said, “It’s an incredibly dangerous precedent to affix somebody’s behavior or certain actions based on emotions or characteristics presented in their face.” In defense to all of the complaints relating to privacy infringement, one school administrator said, “The system only collects students’ facial expressions and behavior information.” Others claim that parents only received screenshots of their child in the classroom, with other students’ faces blurred out.
In the United States, it is anticipated that AI usage in education will grow by 47.5 percent from 2017 to 2021. The Chan-Zuckerberg Initiative and the Bill and Melinda Gates Foundation suggest that further research and investment should be put into using machine learning as an educational tool. Carnegie Learning and Content Technologies are two companies currently developing software that can be personalized based on the struggles of the student using the application. The technology can serve as a supplementary aid to teachers with classes where there are large learning gaps between classmates. With the advancement of AI, algorithms will be sophisticated enough to detect struggles through students’ facial expressions and use that as another meter to calculate skill level. AI would also make it possible for more enriching online classes since it can assist in explaining problems to the students without the physical presence of an instructor. Also, students not fluent in English could use computer applications in another language in order for them to have an efficient system of education.
Not all of China’s AI educational implementations are considered as intrusive as their facial recognition software and headbands. Squirrel AI, an adaptive learning company, currently has centers located around China where students can use personalized educational programs. Unlike the software in the U.S., the adaptive learning systems in China are much more prevalent and easily accessible. Squirrel AI has over 2,000 centers around China. The company’s product is also being used in classrooms to help failing students. Most Squirrel AI users have had a positive experience. One such user, Zhou Yi, had tried the tutoring service and after two years, his math test scores rose from 50 percent to 85 percent. Squirrel AI functions like many other adaptive educational programs. It notes the concepts that the user might not understand and helps them improve on those skills.
AI is a pervasive technology that will only grow in usage throughout the twenty-first century. Its incorporation in the classroom is a scheme that will possibly improve the education of millions of students around the world. Society will need to figure out how to effectively balance the advantages of the technology with their own moral compass and comfort with privacy.
Article by Huan Changvu of Cabin John Middle School
Graphic by Sophie Zheng of Cabin John Middle School