As China’s high-stakes Gaokao 2025 college entrance exams commenced this week, major tech companies including Alibaba and Tencent temporarily disabled key AI functions in their chatbots to prevent potential misuse by students. The move highlights the growing tension between fast-developing artificial intelligence and the need to preserve exam integrity.
AI Restrictions During Exam Hours
Popular Chinese AI assistants such as Alibaba’s Qwen, Tencent’s Yuanbao, and Moonshot’s Kimi suspended photo recognition capabilities during the Gaokao, which is being held from June 7 to June 10. When prompted to identify or solve problems using images, the bots replied with notices like:
While ByteDance’s Doubao and Qwen still displayed photo upload options, both systems rejected exam-related images during testing hours. Qwen, for instance, clarified that the service was frozen for the duration of the Gaokao.
Gaokao: A Critical Exam for Over 13 Million Students
The Gaokao China’s National College Entrance Examination—is widely regarded as one of the most challenging and decisive academic tests in the world. This year, approximately 13.4 million students are sitting for the multi-day exams, which play a central role in determining university admissions and long-term career prospects.
Unlike Western systems where applications consider academic portfolios, personal statements, and extracurriculars, Chinese university entry hinges almost entirely on this single test. The pressure is immense, especially for students from under-resourced backgrounds or rural areas.
A poor score could mean repeating a year of high school, while success can lead to life-changing opportunities.
AI Tools Pose New Cheating Risks
For some time now, an extensive system of rules has been established in China to battle cheating during the Gaokao exams; including prohibiting cell phones, equipping testing locations with cameras, and using signal jamming technology at local exam centers. New concerns now weigh on the shoulders of testing authorities due to the rapid growth of AI technologies.
Tools so advanced exist that they can analyze images of test questions and solve problems, even simulate handwriting; therefore, the potential for AI to disrupt exam security is apparent. Thus, the temporary shutdown of functionality like image recognition could be a very proactive approach in light of that reality.
A Symbol of National Meritocracy
In Chinese society, the Gaokao is considered a meritocratic leveler. A top score on the Gaokao may be the only chance at a top-tier university–and a better life–for millions of students from lower-income families. As a result, protecting the fairness of the exam to both regulators and the public is most important.
With AI now emerging in everyday lives in China, tech firms must increasingly accept the responsibilities that will ensure that their innovations do not erode public institutions like education.
Conclusion
As AI advancements continue to develop and grow, governments farming AI use with various restrictions similar to the way China has used a temporary disabling of features during risky national events could become common practice globally. The intersection of AI and education is emerging as a potential policy minefield for governments, and balancing a progression with limitations and ethics will remain an ongoing dilemma.
About The Author
