Jingli Yan jingliyan@hks.harvard.edu WHEN schools closed early this past winter, online learning platforms immediately promoted courses. All went well except for the ad frenzy that incited a chastising on social media after users spotted the same face featuring different courses in four different apps, raising doubt about their content’s credibility. Education apps are reviewed and approved by the Ministry of Education. In 2019, 11 government agencies issued joint guidelines for online education’s healthy development, Internet promotion, AI and other innovations applicable to education. By the end of 2020, 3,883 apps had been approved. Of the 1,791 companies that have invested in these apps, Baidu, Alibaba and Tencent are well-known tech giants. The huge market and parents’ willingness to pay for education have made online education the next blue sea for China during the pandemic. This past year provided them with a perfect arena to test the water. But they are missing the opportunity to reimagine what really matters in the value of education. With all the technology and hot money, online learning has departed its ways with offline education, leaning towards targeted advertisements and subsidies, and sugarcoating their products in AI and big data. When market competition becomes brutal, courses are offered at meager prices. For example, VIPKID promoted classes as low as US$0.26 to learn English one-on-one from a foreign teacher. A mom of a 1-year-old kid was reported to have signed up for 10 similar courses for less than US$1.50 each last summer. Education is not a free meal. What students need is tailored premium content and an attentive teacher. Both need human capital investment, and neither of them can be created in a rush. Dumping the courses in the market like super sales might fool parents once, but they will become reluctant to pay when they discover their children are not actually learning. “I think the most useful education apps are the free ones that offer instant help with a camera scan,” a mother of a 16-year-old daughter told me. “We tried several other apps but ditched them eventually. My daughter cannot focus on the screen.” When offline sessions resumed, she sent her daughter back to classroom, a relief for both parents and students. Middle school students and parents in China are intimidated by the fierce competition of the university entrance exam. They really don’t have much time to repeatedly test different apps. On the other hand, parents of younger children are more willing to try language apps. A mother in Beijing introduced me to Banma English because lessons are short, asynchronous and cheap. “They also sent out numerous complimentary materials,” she emphasized. Another mom in Changsha chose an offline English training center for her son but acknowledged online learning advertisements had almost captivated her. All those moms feel pressured and anxious, yet they hope children can be motivated to learn. And online learning platforms bragged about their solution: they have exclusive big data on child behavior, and they use AI+big data to match a student to the most effective teacher. However, the result is really opaque. A VIPKID teacher noticed that AI labeled a really chatty Kathy type as reserved. And then, the system recommended that the teacher should encourage her to interact more. When AI falsely identifies whether a child is learning or not, how can the teacher correct it? What if the teacher hasn’t noticed it, or if the platforms use recorded video and AI technique to deliver the course, as is the case in most cheap offerings? AI is a black box churning out the data points that the technologists direct them to collect. According to an engineer at VIPKID, they collect students’ facial recognition data, enrollment and drop-out data to determine the teaching quality. This use of AI seems to be focused on retention and profit rather than on educating a child. The design misses the value of education but is perfectly aligned with the pursuit of capital return. In 2018, Kaifu Lee, a prominent AI expert who is chairman and CEO of venture capital firm Sinovation Ventures, imagined that online-merge-offline education would ride the third wave of AI perception and be a game-changer after AI replaces the doctor, judge and banker. Two years later, he would be surprised to see that education apps replaced teachers of different courses with the same face. Online platforms need to use this opportunity to reimagine what really matters to users. If they are not successful selling courses like group-buying cabbages with a one-face-fits-all advertisement, perhaps they should reflect on the fact that they all choose the same image of a highly sought-after, experienced teacher. And that’s exactly what AI and big data cannot replace in online education. (The author is a mid-career MPA student at Harvard Kennedy School. She briefly worked at Yunnan University of Finance and Economics as a lecturer and has almost 10 years’ public policy experience focusing on international development.) | 