KisKIan
Children write - KI answers

Dr. Maurice Fürstenberg

The research project Kinder schreiben – KI antwortet (KisKIan) investigates the quality of feedback on student texts provided by both human teachers and artificial intelligence systems. The aim is to improve the quality of such feedback in order to promote students' writing skills. At the beginning of the project, the focus will be on analyzing the quality of machine feedback, which has hardly been researched in German-speaking countries to date. To this end, criteria-based feedback from artificial and human intelligences will be compared. In the long term, the aim is to develop methods for improving machine feedback on student texts, such as targeted top-down prompting using predefined assessment criteria. To achieve this, however, a common sense of assessment criteria for different text types and age groups must first be developed, which experts (teachers and didacticians) consider to be a solid basis for effective feedback. A cooperation with the feedback platform fiete.ai and workshops with experienced teachers will help to condense these criteria, as will conferences with German didactics experts and the resulting anthologies.

Further information and the latest status of the project can be found here.