Teachers’ Perceptions of AI-Generated Feedback in EFL Writing : A Study of Swedish EFL Teachers’ Evaluations of AI-Generated Corrective Feedback and Formative Feedback in Years 4–6

Sammanfattning

This study investigates the use of artificial intelligence (AI) as a tool for providing corrective and formative feedback on written English texts produced by EFL learners in Swedish compulsory school (Years 4–6). Using a mixed-methods comparative design, the study explores both the effectiveness of AI-generated feedback and certified English teachers’ attitudes toward its use in assessment practices. Data were collected through two digital surveys with 118 teachers and two semi-structured interviews. One survey group was informed that the feedback had been generated by AI, while the second group evaluated the same feedback without this information. The findings show that teachers generally viewed the AI-generated feedback as useful for grammar correction, vocabulary development, and formative assessment. However, teachers who were aware that AI had generated the feedback evaluated it more critically than those who were unaware of its origin, indicating tendencies toward algorithm aversion. The study concludes that AI has significant potential as a complementary support tool in English language education, particularly for providing immediate formative feedback. At the same time, successful implementation depends on teacher involvement, professional development, and ethical awareness.