Authorities in the Chinese region recently announced the launch, in the next academic year, of “a localized service platform within the scope of AI education.” This tool will include functions such as “intelligent statement composition, intelligent correction, and other features, aimed at reducing teachers’ workload, from lesson preparation to assessment.”
Although supportive of the platform, teachers stressed to Lusa the current limitations of the technology. “Using AI to grade assignments can result in errors that require manual correction,” said Ruan Zhanpeng, an information technology teacher at Pui Va Secondary School.
Ruan acknowledged that AI can reduce workload in “marking simple multiple-choice questions, but for open-ended questions, manual adjustments are still necessary.” Chinese language teacher Nora Lam, from the Macau Residents’ School, shared the same view regarding AI’s limitations in handling questions that require development.
“Review is necessary after AI corrects a composition, because it cannot understand texts based on feelings,” she noted. Pedro Lobo, a teacher with more than 30 years of experience in information technology in the territory, revealed that he has used AI as “a starting point in preparing lessons and materials for students.”
The Portuguese educator agreed it can be a useful tool for teachers but emphasized the need for training.
For teachers who don’t speak Chinese, I have hardly seen any training,” he told Lusa.
The Education and Youth Development Bureau (DSEDJ) stated that, for the current academic year, it provides at least 10 hours of training for information technology teachers and at least six hours for teachers of other subjects.
“In the 2025/2026 academic year, around 3,000 training slots and more than 51,800 teaching hours were offered to IT teachers, and around 9,000 training slots and more than 355,000 teaching hours to teachers of other subjects,” the DSEDJ said in a written reply to legislator Lei Cheng I regarding strategies to reduce teachers’ workload.
Platform with Lusa
