0
First and second-year students at the High School Affiliated to Luoyang Institute of Science and Technology in Luoyang, Henan Province, throw paper airplanes on May 27 to cheer on third-year seniors ahead of the upcoming Gaokao. / Photo via The Beijing News

China Central Television (CCTV) reported on May 29 that major artificial intelligence platforms across the country have confirmed that while their systems will remain operational during the upcoming national college entrance examination, or Gaokao, functions capable of assisting with exam queries will be placed under strict controls. In tandem with mainstream AI applications, online homework-assistance portals targeting primary and secondary school students will also face heavy restrictions on their active question-and-answer capabilities, the state broadcaster added.

The functional limits imposed by these AI platforms are projected to follow a highly synchronized timeframe. This structural approach means that utilities designed for photograph-based question recognition, prompt processing, and automated answer explanations will be temporarily disabled exclusively during active Gaokao testing windows. Conversely, routine conversational features addressing everyday lifestyle inquiries will remain fully accessible without technical constraints.

According to industry sources, multiple domestic AI providers previously implemented similar technical limitations during last year's examination cycle to comply with regulatory mandates. However, the enforcement metrics have been significantly elevated for the 2026 cycle. The stricter security compliance follows a joint video conference convened last month by the Ministry of Education, the Cyberspace Administration of China (CAC), and the Ministry of Public Security. During the session, the interagency coalition demanded higher operational benchmarks for examination security and explicitly ordered a comprehensive crackdown on high-tech cheating methods.

CCTV noted that the stringent regulatory measures have triggered a polarized debate across Chinese digital networks. A segment of netizens questioned whether capping generative AI utilities is an overreaction, arguing that current high-stakes testing facilities are already subject to military-grade surveillance and physical security checks. Nevertheless, the vast majority of online commentators supported the pre-emptive measures, countering that traditional proctoring methodologies face systemic vulnerabilities from microscopic smart devices and the sub-second processing speeds of modern large language models. Proponents argued that shutting down algorithmic pipelines at the source is an inevitable necessity to guarantee absolute fairness across the national educational system.

                                                                                                            Hong Soon-do