The Future of Education

Teacher assessment competence – understood as the set of knowledge, skills, and practices that enable teachers to design, implement, and interpret assessment effectively – represents a fundamental professional competence (Yang, 2024). The use of an AI-based toolkit modifies teachers' assessment practices (differentiation, feedback, accessibility), with variations across school levels (Bower and Torrington, 2024; Isidori et All., 2025a).

We analyzed 62 teaching projects (preschool n=18, primary n=16, lower secondary n=14, upper secondary n=14), developed by teachers working in subgroups. Teachers integrated an AI toolkit (LLMs, generative platforms, interactive quizzes). Four parameters were measured: differentiated assessment, formative feedback, voice recordings (accessibility), and explicit AI documentation. Impact was assessed using a structured coding scheme (presence/absence). The distribution was balanced; light statistical weighting and sensitivity analysis were applied.

Secondary school teachers showed marked resistance to applying the toolkit. After weighting:

• Differentiation: 72% in preschool/primary vs 21% in secondary.
• Formative feedback: 38% in primary vs 14% in secondary.
• Multimodal accessibility: 56% in preschool vs 4% in secondary.

Reported resistances include: distrust in AI, preference for traditional assessment, and the perception that accessibility is unnecessary in secondary school.

AI can redesign assessment competence, but secondary teachers' resistance limits its impact. Professional development pathways must address cultural and motivational barriers.