OECD survey: one in three lower-secondary teachers use AI, three in four lack teaching skills

The OECD's TALIS 2024 survey found that about one in three lower-secondary teachers in surveyed education systems used AI in the previous year, while three in four said they lacked the skills to teach with it. Among users, 68 per cent used AI to learn about or summarise topics and 64 per cent to prepare lessons, while only 25 per cent reviewed student data and 26 per cent assessed or graded work. Nearly three-quarters of teachers in Singapore and the UAE reported AI use, compared with fewer than one in five in France and Japan.

The survey provides a verifiable teacher-side baseline for AI education policy: use clusters around efficiency tasks such as summarising and lesson planning, while grading and data review remain less common, indicating efficiency support rather than automated teaching. Self-reported teacher data is not a measure of student learning, but with three in four teachers saying they lack skills, training limited to prompt-writing would not cover verification, classroom practice and safeguarding. Before procuring AI tools, schools need to clarify whether teachers can correct AI feedback and how pupils disclose AI assistance; otherwise rising usage will not automatically improve teaching capacity. If national policies push AI curricula without teacher training, devices, local-language resources and privacy rules, they risk widening rather than narrowing existing gaps.