Given that developments in AI education have been spearheaded by the private sector, one of the most significant challenges is whether or not AI technologies can be widely implemented in the existing Chinese public school system. To continue the ride-sharing analogy, although Uber was able to integrate itself into existing public transportation systems in some countries, elsewhere the company’s presence disrupted the local taxi cab industry. Public school systems do not operate at the speed of startups – they are embedded in long-standing bureaucracies, local authorities driven by their own interests and teachers unfamiliar with novel technologies. In order for new innovations to achieve widespread impact, public schools will need to collaborate with private companies and ensure that these collaborations are fundamentally driven by student interests, not maximising the company’s profits.
Moreover, for these innovations to have impact, they must be made widely accessible. China’s education system remains deeply divided, often between wealthy schools in the nation’s first-tier cities and poorer schools in its rural regions. Schools that have the ability to introduce innovative new pedagogies are most likely the nation’s elite.
A 2019 report by the Mercator Institute for China Studies on creativity in Chinese schools points to a widening gap between China’s poorer and more prosperous regions. Wealthy schools ‘offer a variety of novel teaching approaches, like maker spaces, escape rooms and advanced computer classes … [Their poorer counterparts] are struggling with limited resources and are struggling to establish creativity-fostering environments.’ Some companies, such Squirrel AI, have diverted their attention away from China’s urban centres, focusing their efforts on opening over 1700 schools in the nation’s second and third-tier cities. Nonetheless, relying on the private sector alone may perpetuate rather than narrow existing inequities within the education system.
Furthermore, while intelligent technologies may be capable of teaching skills that are easily quantifiable, such as middle-school mathematics, we do not know whether they can teach more complex, soft skills such as creativity, critical thinking and collaboration. ‘Intelligent technologies are useful for the rapid acquisition of knowledge,’ says Professor Yong Zhao, University of Kansas. ‘But does that amount to a higher quality of education?’ China’s current high-school education system – rigidly standardised, outcome driven and metrics based – revolves entirely around preparing students for the gaokao, its national university entrance exam, and is criticised for its excessive emphasis on test scores and rote memorisation. In this system, students are still prepared for the workplace of the industrial age, treated as passive recipients of knowledge transfer, like minds on a conveyor belt. If tailored for the current system, intelligent tutoring systems may simply bolster the existing gaokao education system and train students to become better test-takers.
However, as Rogier Creemers highlights in his essay in this collection, there are limitations to the ‘engineering’ approach when it comes to those areas of life that are more complex and unpredictable. If China were to define a high-quality education as one that cultivates creative and collaborative critical thinkers, capable of contributing meaningfully to society, what role can intelligent technologies play in teaching the next generation of students? Can an adaptive learning system teach students how to cultivate supportive, interpersonal relationships – crucial to effective collaboration? Can a robot tutor teach emotional resilience – a skill critical to navigating the uncertainty of the 21st century? Can a video tutorial show students how to make meaning from their own experiences and learn from their mistakes? Even if there existed technology that could successfully teach these skills (for example, Squirrel AI is collaborating with Stanford University to research AI applications in the teaching of leadership and creativity), would it be in the interests of China’s one-party government to implement such programmes in the nation’s public schools?
Finally, any implementation of AI technology raises the question of the ethical use and collection of data. Students, parents and educators should be informed of the data being collected and how it is being used, companies should articulate transparent best practices and governments should establish clear regulations. What data is helpful and what is harmful to a student’s educational experience? On one hand, data is a crucial asset that allows for intelligent technologies to deliver feedback and address student needs with much greater precision and efficacy. An education company might collect information on the time it took a student to answer a question to better evaluate their abilities. On the other hand, data could be used as a means of maximising a company’s profit. Personal identifiable information – say, a student’s sexual orientation or biometric data – could be used to discriminate or target a student on the basis of their sexuality or physical experience. Reports even show some schools deploying AI-powered gates and facial recognition cameras to monitor everything from student concentration levels to their emotional state.
Whereas the human teacher assumes change, AI assumes continuation. These are fundamentally different approaches to an individual's capacity for change and growth
Although the Chinese government has taken some steps to 'curb and regulate' the use of facial recognition and new technologies in schools, without stringent regulations these technologies have serious implications for student privacy, as Danit Gal discusses in her essay. If schools are capable of tracking every keystroke, knowledge point and facial twitch, they are effectively furnishing either a technology company or the Chinese state with an eternal ledger of every step of a child’s development. This is potentially problematic because, whereas the human teacher assumes change, AI assumes continuation. Today, when a kindergarten student makes a mistake, the human teacher will try to help the student overcome it and the mistake will eventually be forgotten. In contrast, an intelligent tutoring system could not only store that information and tailor a personalised pathway for the student in the first grade, it may extrapolate that information many years later, when the student is in high school.
These are fundamentally different approaches to an individual’s capacity for change and growth. Should schools use your behaviour in high school to judge or predict your trustworthiness as an adult? Should colleges have access to a student’s genetic predisposition for mental illness, when sorting them into dormitories? If algorithms can identify pre-existing signs of prodigy and talent, can schools pre-emptively recruit them into special programmes? These are all questions that must be grappled with when the task of evaluating student potential has been handed over to an algorithm.
