TLDR: Major Chinese artificial intelligence companies, including Alibaba, ByteDance, DeepSeek, Moonshot, and Tencent, temporarily suspended certain AI chatbot functionalities, particularly image recognition, from June 7 to June 10, 2025. This unprecedented measure was implemented during the annual Gaokao, China’s national college entrance examination, to proactively prevent students from using AI tools for cheating and to ensure the integrity and fairness of the high-stakes exams.
In a significant move to uphold academic integrity, leading Chinese artificial intelligence firms, such as Alibaba, ByteDance, DeepSeek, Moonshot, and Tencent, temporarily deactivated key features of their AI chatbots during the nation’s rigorous Gaokao (National College Entrance Examination) from June 7 to June 10, 2025. This widespread suspension, which affected popular tools like Alibaba’s Qwen, ByteDance’s Doubao, DeepSeek, Moonshot’s Kimi, and Tencent’s Yuanbao, was a direct response to mounting concerns over the potential misuse of generative AI for exam cheating.
The primary functionality targeted for suspension was image recognition, which could potentially allow students to upload photos of exam papers and receive instant answers. When users attempted to access these features during the restricted period, chatbots like Yuanbao and Kimi explicitly stated that the functions were deactivated ‘to ensure the fairness of the college entrance examinations.’ Doubao also indicated that uploaded pictures were ‘not in compliance with rules,’ while DeepSeek restricted its service during specific hours for the same reason.
This proactive measure was part of a broader, comprehensive effort by Chinese authorities to prevent cheating in the Gaokao, an exam critical for the future academic and professional paths of over 13.3 million students. Beyond the AI chatbot suspensions, other stringent anti-cheating methods employed included biometric verifications, signal jamming to disrupt electronic communications, strict filtering of digital devices, and extensive video surveillance. AI technology itself was also utilized to monitor student behavior during the exams, looking for suspicious actions like whispering or repeated glancing, with flagged footage reviewed post-exam.
Also Read:
- Cybersecurity Leaders Urge Immediate AI Regulation Amid DeepSeek Concerns
- Canadian Universities Embrace AI Tools Amidst Persistent Concerns
While the increasing accessibility of AI chatbots has raised global concerns about academic dishonesty, China’s centralized control over its tech landscape allowed for such a coordinated, nationwide suspension. In contrast, similar efforts to prevent AI-assisted cheating in high-stakes exams have not been widely observed in countries like the United States, where the educational system is more decentralized and government control over tech firms is less direct. It is important to note that no actual cheating incidents directly caused by AI were reported; rather, the disabling of these AI features was a preventive measure to safeguard the fairness and integrity of the Gaokao.
