Artificial intelligence is arriving in classrooms faster than many teachers can keep up with, and for some it brings stress as well as new possibilities. A new study from China suggests that teachers who feel confident using AI across every part of their job tend to enjoy their work more, although anxiety about the technology can take the edge off that benefit.

Researchers at Jinhua University of Vocational Technology surveyed 768 vocational education teachers across eastern, central and western China. The findings appear in BMC Psychology. Roughly half worked in secondary vocational schools and half in higher vocational colleges, where staff combine classroom teaching with practical training.

Teachers rated their confidence in six areas of AI teaching, from technical skills and lesson design to assessment and the ethics of using AI. They also reported how anxious AI made them feel and completed a questionnaire on occupational wellbeing, which reflects how satisfied, engaged and positive people feel about their work.

Rather than ranking teachers from low to high, the team sorted them into groups with similar patterns. About one in five were at a foundational stage, with low confidence across the board. Another fifth were technically skilled but cautious about ethics, while about a quarter were confident in teaching methods but held back by technical skills. The largest group, nearly a third, felt confident in every area.

Wellbeing rose steadily across these groups, from the foundational stage to the all round confident group, and the difference between the two ends was large. Even the two middle groups differed, with the teaching focused group reporting higher wellbeing than the technically skilled one.

Teachers who used AI tools more often and had attended more AI training sessions were more likely to belong to the confident groups. Gender, age, teaching experience and subject made no clear difference. AI anxiety reduced the wellbeing advantage of the most confident group by roughly a third, although its overall effect was small.

The findings suggest that one size fits all AI training may miss the mark. Teachers at the foundational stage may need broad basic support, while technically skilled teachers may gain more from guidance on ethics and teaching than from yet more technical courses. Tackling anxiety about AI could help confident teachers get the full benefit.

The study has limits. It used a single online survey, so it cannot show that confidence with AI causes better wellbeing, and all answers were self reported. Schools were chosen through cooperation rather than at random, and the results may not apply to teachers outside vocational education or outside China.

Governments around the world are pushing schools and colleges to adopt AI. This study suggests that supporting teachers’ confidence, and their peace of mind, may be just as important as supplying the technology.