99.7% of Chinese Students Have Used AI, Survey Finds

A major new survey by the China National Academy of Educational Sciences has found that nearly 99.7% of primary and secondary school students in central and western China have been exposed to artificial intelligence, with 85.6% using AI to complete homework assignments. The findings, presented at Beijing Normal University’s inaugural “Jingshi Think Tank Dialogue” on May 8 and published by Guangming Daily on May 22, reveal a stark disconnect between student behavior and adult awareness that experts say poses urgent questions for the future of education.

The Awareness Gap

The survey, conducted during summer 2025 across all 31 provinces, collected data from 320,000 students, 280,000 parents, and 50,000 teachers. While 99.7% of students reported AI exposure, only 13.8% of teachers and 17.8% of parents believed that students commonly use AI — a gap that Li Yongzhi, President of the China National Academy of Educational Sciences, described as a “cognitive blind spot.”

“Teachers and parents have a clear cognitive blind spot,” Li said at the dialogue, as reported by People’s Daily. “This shows that a large number of students are using AI in an unsupervised, unguided state.”

Li warned that without proper guidance, students risk “ceding cognitive sovereignty” — voluntarily giving up independent thinking in favor of AI-generated answers.

A Tale of Two Users

The research reveals a sharp polarization in how students engage with AI. Self-disciplined students use AI as a cognitive expansion tool, while those with weaker self-control treat it as a crutch for completing assignments. Crucially, the determining factor is not family resources or school quality, but internal drive.

“The core variable causing the divide has never been family or school resources, but internal drive,” Li explained. “AI will never actively think about ‘what is meaningful to do,’ but humans can.”

This finding reframes the debate from one of access inequality to one of motivation and self-regulation, suggesting that educational interventions should focus on cultivating intrinsic motivation rather than simply restricting AI access.

The Urban-Rural Dimension

Data from a separate but related survey by the China Youth and Children Research Center, published by CCTV in March 2026, adds a geographic dimension to the findings. That survey of 8,563 students across seven provinces found that over 60% of primary and secondary students had used generative AI, with 71% using it for homework assistance.

Rural students use AI for homework at higher rates (73.2%) than urban students (68.4%), but also show higher rates of AI dependency — 22.8% of rural students said they want to rely on AI rather than think for themselves, compared to 17.7% of urban students. Overall, 20.5% of surveyed students expressed a desire to rely on AI for thinking rather than thinking independently.

What Skills Matter in the AI Era?

The Jingshi Think Tank Dialogue brought together leading educators to address a fundamental question: what should education cultivate when AI can perform so many cognitive tasks?

Zhao Yong, Distinguished Professor at the University of Kansas and Academician of the American Academy of Educational Sciences, argued that the most essential skill is the ability to discover and solve problems.

“AI can provide countless standard answers, but it can never raise warm, valuable real questions,” Zhao said, as reported by China Daily. He emphasized that judging the value of a problem requires asking: Why does this matter? Why am I qualified to solve it? Why now? These questions, he argued, are far more important than solving a predetermined problem.

Zhou Zuoyu, Professor at Beijing Normal University’s Faculty of Education, noted that human society has entered a new pattern of “human + AI” collaboration. Education must break free from the traditional logic of cultivating a single ability, he said, while remaining vigilant about the risk that AI could widen social divides.

Policy Implications

The findings are likely to influence Chinese education policy in several ways. Experts at the dialogue called for AI literacy education for both students and parents, the development of guidelines for appropriate AI use in schools, and a renewed emphasis on cultivating “core competencies” that AI cannot replicate — problem-finding, critical thinking, and ethical judgment.

China has been aggressively integrating AI into its education system as part of the State Council’s “New Generation Artificial Intelligence Development Plan” (2017), which set goals for AI education at all levels. The China National Academy of Educational Sciences publishes an annual “China Smart Education Development Report,” with the 2024-2025 edition released in February 2026.

What to Watch For

As AI becomes nearly ubiquitous in Chinese students’ lives, the central challenge for educators will be guiding its use toward cognitive enhancement rather than dependency. The research suggests that the key battleground is not access — which is already near-universal — but the cultivation of the internal drive that determines whether AI becomes a tool for growth or a crutch that weakens independent thought.

Li Yongzhi put it succinctly: “The most urgent thing to be vigilant about is humans voluntarily giving up thinking and ceding cognitive ability. This issue needs to be led and resolved through educational reform.”