EdTech and China’s Left-Behind Kids
China has 69 million “left behind” rural children. They are left behind literally because their parents work as migrants in wealthy cities, which do not permit non-residents to attend public schools. They are also left behind figuratively because as China’s economy roars ahead their future is uncertain. Rural schools do not attract good teachers. And as more families move to urban centers, rural schools are forced to close down.
The Chinese government believes that EdTech is the solution. Today, over 90 percent of rural areas have broadband access, and urban teachers can livestream their lessons into rural schools.
The results have been disappointing.
The main issue is that the content produced in urban classrooms does not meet the needs of rural ones. But there is a solution on the horizon: Artificial Intelligence.
In the past decade, AI has changed the way we live and work by identifying the hidden patterns interweaving human society, and using user feedback to customize the way we connect and buy (think of the Amazon recommendation engine). While the theories underlying AI have been with us for decades, it is the confluence of three forces that has enabled AI to become an everyday technology.
- Moore’s Law, which states that the processing power of a micro-chip will double every two years. This has enabled the rise of Big Data.
- The prevalence of smartphones, which permits billions to access and interact with the internet. This has allowed the internet to collect data on our habits and preferences.
- The rise of Big Tech, which means that all this Big Data is centralized and consolidated. That’s why the leading AI companies are Google, Facebook, and Amazon in the United States, and Alibaba, Baidu, and Tencent in China.
AI is most readily achieved in a centralized system that allows for Big Data collection, and has immediate feedback loops. Of all the school systems in the world, Chinese education is the most AI-friendly because:
- China has a cultural obsession with test scores. Because schools only care about test scores, this allows for limited parameters and a defined output, which is crucial for AI. Tests also provide a feedback loop to train and test the AI system.
- China has a different concept of privacy. This allows any and all data to be collected. Chinese schools have used facial recognition technology and brain scans to analyze students’ attention levels.
- China has a centralized curriculum, which allows for regularity, homogeneity, and scale in data collection. From grade one to university, fifty students sit in a classroom, and are lectured to. The static and repetitive nature of Chinese pedagogy permits it to be mathematically expressed, and analyzed by the AI system.
While Chinese policy-makers have wanted to implement AI in education for years, EdTech has always been too fragmented for AI purposes. The COVID-19 crisis offers a rare opportunity for Beijing to step in and consolidate the various EdTech platforms controlled by local governments.
AI in education promises to help “left behind” children, while at the same time it also risks further alienating them. The success of AI in China’s education system is highly dependent on the approach China’s government takes in introducing it into the system.
China’s Choice
An AI system in education can focus either on improving test scores, or improving classroom teaching.
Improving test scores seems the easier choice, with the most immediate benefit. Imagine a system in which students are doing test questions all the time. With its database of every test question ever written, the AI system can figure out students’ knowledge gaps, and then pepper students with customized tests. They will not understand the material, but their test scores will go up, and that’s all that matters.
Such a system would mean that Chinese education becomes “slave to the algorithm,” and teachers lose their professional autonomy, becoming nothing more than ICT assistants.
This will mean that China’s “left behind” children become truly so.
To understand why, let’s talk about the 1995 book Meaningful Differences by the two American researchers Betty Hart and Todd Risley. For over two years, they recorded the words spoken at home in professional and poor households, and coined the phrase “30 million-word gap by age three.”
A technologist would think there’s an easy solution here: design a robot to converse with a poor child.
Hart and Risley argue that would be the wrong approach. It doesn’t matter how many words are said. What matters is how words are said. Poor parents treat their children as subordinates: “Don’t touch that!” Rich parents treat their children as equals: “Let me explain why touching that is dangerous.” That’s why Hart and Risley think poor parents should play board games with their children.
Let’s look at a classroom example of this idea. In 1968, in response to the assassination of Martin Luther King Jr., an American third-grade teacher named Jane Elliott conducted an experiment to teach her white students about the corrosive effects of discrimination. For one week, students with brown eyes sat at the front of the class, and received endless praise. Students without brown eyes sat at the back, and received harsh criticism. At the end of the week, Elliott discovered that the brown-eyed students became more motivated to learn, and the other students became less so.
As a teacher trainer, I encourage teachers to recognize their hidden biases, and build emotional connections with students. I tell them to always keep calm, smile at every student once a day, and laugh as much as possible.
An effective AI system could test and verify these hypotheses, and assess the impact of emotional connectedness on test performance. It could spot patterns that are invisible to the human eye. Does regular eye contact affect motivation? Does tone of voice? What if every student has a chance to speak up in class?
Emotions are the real information superhighway, and it is emotional connectedness that undergirds any effective EdTech system.
I consult Chengdu Experimental Primary School and Chengdu #3 Kindergarten, two flagship public schools that connect to hundreds of schools in poor rural areas. Their teachers livestream their classes, email lesson plans to partner schools, and offer feedback over instant messaging.
This system works because teachers visit the rural schools once a semester and become friends with the teachers they train. The rural teachers are invited to visit Chengdu often and sit in on classes. Technology has not replaced teachers—it has helped teachers build a collaborative support community.
An AI system, no matter how effective, cannot change the destinies of China’s “left behind” children by itself. What’s really needed are progressive government policies that help these children emotionally connect into a supportive community.
China will use COVID-19 to embrace AI in education. How China does so will determine the future of its children, and perhaps even the future of education itself.