Beyond Grades: How AI Can Help Us See Our Students
Petra Stehr, Oberschule Ofenerdiek
When people hear “artificial intelligence in schools”, they often think of faster grading or auto‑generated worksheets. I see something very different. At Oberschule Ofenerdiek, where I teach Art, Music, German and Biology and work as a counsellor, I experience every day how much more AI could do to help us truly understand and support our students.
My work – both at school and as a creative and trauma therapist – constantly reminds me that children don’t just need food, safety and a warm classroom. They also need attachment, competence and autonomy if they are to develop in a healthy way. When these psychological needs are not met – because of family crises, mental health issues, trauma or undiagnosed learning and developmental difficulties – this often appears in school as aggression, withdrawal, disinterest, absenteeism, anxiety or loss of motivation. Behind every “difficult” student I meet, there is an unmet need and a story that has not yet been heard.
For me, EduAId is an opportunity to rethink how we respond to these stories with the help of AI. One of my central ideas is to build a regional competence centre for everyone working with children and young people in schools. In such a centre, AI could support us by collecting cases, anonymising them, and helping us recognise patterns across schools: what helps students with ADHD stay engaged, which structures support autistic learners, how trauma shows up in classroom behaviour. This shared knowledge base would not replace professional judgement, but it would give teachers and counsellors a richer, data‑informed starting point.
At the same time, I believe in the importance of regular teacher professional development – what we call Schilf in Germany. Here, AI could curate up‑to‑date research on mental health, developmental and attachment disorders, and suggest tailored training modules based on the needs of a particular school. Instead of every staff member searching alone, we could have AI‑supported workshops, simulations and case discussions that speak directly to our everyday challenges: the student who suddenly stops coming to school, the quiet child who never speaks, the teenager whose anger fills the classroom.
In my own practice, I imagine AI‑supported questionnaires for students as a powerful tool. Children and young people could describe, in their own words and at their own pace, how they feel at school, what stresses them, when they feel safe, and when they feel invisible. AI can help analyse and visualise these responses so that I, my colleagues and school leaders see more than just grades and attendance records; we see emotional climates, warning signs and also hidden strengths. This kind of insight allows us to adapt our approach before problems explode.
Another idea that is close to my heart is the creation of AI‑driven “simulations” to foster empathy. I imagine short, interactive experiences where colleagues and classmates can briefly “step into the shoes” of a student living with blindness, autism, anxiety, trauma, or sensory overload. Guided by AI, these simulations could make it easier to feel, not just know, why a certain student avoids eye contact, reacts strongly to noise, or seems “lazy” when in fact they are overwhelmed. For me, such perspective‑taking is one of the most powerful starting points for changing classroom routines, expectations and support strategies.
EduAId also opens up creative space in everyday teaching. With AI, I can design differentiated learning materials for students with dyslexia, dyscalculia, ADHD, autism or a migration background much more efficiently. Instead of preparing several versions of a worksheet late at night, I can co‑create them with an AI tool and then adapt them with my professional judgement. This means more energy left for what I value most: building real, trusting relationships with my students. At the same time, AI can help us rethink learning environments themselves – from quieter corners for anxious students to activity‑based options for those who struggle to sit still.
I also see great potential in AI‑supported peer‑to‑peer exchange. Within EduAId, schools could share anonymised experiences through intelligent questionnaires: what works in one context, where we struggle, which small interventions have made a big difference. AI can cluster and highlight recurring themes – for example, how schools support students with a migration background or how they manage stress on exam days – and feed these insights back into training and school development. All these ideas are driven by one conviction: children and young people with diverse needs should not have to change to “fit” the system. We have to redesign our learning environments with a deeper understanding of their needs. Used creatively and responsibly, AI will never replace the human relationships I build with my students – but it can help me, and many other European educators, to see more, understand more and respond with greater empathy to the young people in our care. EduAId, for me, is not about making teaching more efficient; it is about making schools more humane with the support of intelligent tools.
