Introduction
Selin Aras first noticed the classroom had changed when three homework essays arrived with the same polite rhythm and none of the usual odd turns of language. The problem was not only cheating. It was that the old assignments no longer asked the right questions.
Selin lives in Stuttgart and teaches in a mixed comprehensive school. She now develops AI teaching materials that treat students neither as suspects nor as passive consumers of tools.
Story of the Path into AI
At first the staff room split into two camps. Some wanted to ban AI; others wanted to automate half the workload. Selin found both answers too easy. Her students used the tools to copy, yes, but also to understand, translate and practise. The task was to make thinking visible again.
She learned the basics of prompting, privacy and bias, then redesigned assignments. Instead of asking only for a final text, she asked students to compare AI answers, mark errors, check sources and write short reflections on what they changed. Her first unit had too many rules and too little breathing room. Students spent more time asking what was allowed than thinking about the content. Selin cut the rules down and made the criteria clearer.
Current Work
Today Selin advises schools on AI policies and creates materials for heterogeneous classes. In one literature lesson she had an AI draft a character analysis. Students then marked gaps, false claims and missing textual evidence. The discussion moved from prohibition to responsibility, language and their own work.
Selin is careful not to romanticize the situation. Some students use AI to avoid effort; others use it because school language has locked them out for years. Fair materials must make both realities visible. A blanket ban would miss half the problem.
Personal Advice
“Bans do not explain the world. Good tasks make it negotiable,” Selin says. She advises teachers to assess process, revision and judgement, not just polished final products.
Key Facts
Age and place: 30, Stuttgart.
Background: young teacher, diverse classes, institutional disagreement.
Entry into AI: teaching unit in which students correct AI answers and check sources.
Focus today: school-based AI literacy.
Typical tools: language models, source checking, lesson design.
Werkstattnotiz
Selin keeps one AI essay that is almost good enough. It is useful because the error is small and hidden in the interpretation. Her next unit begins with that text. Students have to prove not that it was generated, but where it stops reading carefully.