Top Chinese AI firms went a step further and suspended major operations of their popular chatbots temporarily to stop students from cheating in the nation’s most eagerly awaited annual academic examination.
While over 13.3 million Chinese students endure the gaokao college entrance exams between June 7-10, technology giants such as Alibaba, ByteDance, and Tencent have disabled or limited their AI products’ strongest functions.
The act serves to raise increasing fears that how readily available AI technology can compromise the integrity of education during high-stakes exams.
AI Giants Suspend Features to Combat Cheating in China’s Gaokao Exams
Alibaba’s Qwen and Doubao by ByteDance have especially suspended their image recognition features from answering questions about test documents, according to Bloomberg. Tencent’s Yuanbao and Kimi by Moonshot have gone a step ahead and suspended their photo-recognition features entirely during the exam period.
The restrictions extend beyond the biggest players. DeepSeek, the foreign artificial intelligence service that was in the news this spring, is also shutting down its features for parts of the day. When students try to use those features, they get messages saying functions have been turned off “to ensure the fairness of the college entrance examinations.”
This collective push by Chinese AI vendors is a testament to the enormous pressure that comes with the gaokao exams.
Contrary to the majority of Western educations that account for other factors in admission to university, the gaokao is the only route of entry to China’s universities. Students prepare for several years to sit for these multi-day exams, aware that their marks will control much of their school and professional life ahead.
The intense competition has created a scenario where any edge on offer may be tempting. While students are already prohibited from carrying phones, laptops, and other gadgets into the exam rooms, the self-regulatory measures of the AI companies add a further deterrent to any attempts at cheating.
What is significant about this situation is how it came to be. According to the Guardian, word of the shutdowns wasn’t issued in formal company announcements but rather disseminated by student chatter on Weibo, China’s social media site of choice.
That this is grassroots discovery means students were attempting to use these AI tools actively and found the blocks for themselves.
How AI is Reshaping Academic Integrity and Testing?
The trend is not exclusive to China. Schools globally are trying to understand how AI chatbots are transforming the landscape of academic honesty.
The Wall Street Journal in May reported that American colleges and universities have experienced a resurgence for old-fashioned blue books—the paper booklets where exams are taken by hand—as schools seek to fight AI-assisted cheating by moving back to analog test-taking methods.
This movement towards paper-based testing is a fascinating reversal of the adoption of education technology. As schools embraced digital technologies and online testing, the emergence of sophisticated AI has prompted many to reconsider the security of digital forms.
Chinese businesses’ self-initiated cuts in services during gaokao are an expression of social responsibility beyond the legal framework. While there may appear to be no such laws mandating the shutdowns, the businesses seem to be aware of their role in educational equity.
The short-term limitations also show how AI companies are trying to weigh innovation against social responsibility. Rather than blacking out their capabilities in the long term, they’re creating separate blackout time frames to coincide with peak learning periods.
As AI technology spreads and becomes more widespread, this Chinese model can serve as a template for other countries facing the same challenges. Whether or not AI will influence education is not the question—the fact that it is already doing so is not in dispute. The question is how to capitalize on its benefits without compromising the integrity of academic testing.
The ban on gaokao is just one example of society’s adaptation to the revolution induced by AI, setting new standards and assumptions for when and how such powerful tools are employed.
