Major Chinese AI platforms temporarily disabled image creation and recognition functions during the gaokao — the most important exam for university admissions. On this day, 13 million students face a test that could change their lives, and no artificial intelligence is prepared to assist. Doubao, DeepSeek, Qwen, Yuanbao, Kimi—all have stated that they are currently “not available for use.”
And what does this imply?
For the first time, the education system acknowledges that it cannot cope with the situation. Knowledge assessments are losing their meaning when AI knows everything and is always at hand. China’s solution is simply to turn off these technologies temporarily. However, this is not a true fix—more of an admission that the system is unable to face reality. As a result, in the educational sphere, if you’re involved in edtech, consulting, or training, you might consider that traditional testing has lost its relevance. Replacing them are models that no longer serve as validators but are more about supporting the process. Learning becomes supportive rather than strict selection.
A school where classes are conducted under AI supervision to later, at the same time, avoid using it—this is like a metaphor for our times. People are pondering what to do next, as coping with challenges in today’s reality grows increasingly difficult.