Honest Feedback

The competence of giving oral feedback in different scenarios, notably to pupils, colleagues and parents, is necessary for starting teachers. Starting teachers state that they are ill-prepared for their workplace since any training they receive in feedback relates to solely peer-feedback, which is limited in relation to their professional practice. Consequently, many starting teachers feel insecure and isolated , which increasingly leads to early resignations , hence contributing to the growing problem of teacher-shortages , leaving a practical gap. An AI-driven Avatar as critical trainer, adaptable to personalised scenarios, could assist in feedback training.
As AI technology increasingly becomes self-educating, a disregard for human values may inhibit interaction with the technology, causing starting teachers to be wary of privacy and autonomy , among others. These human values should therefore be structurally synthesised into AI-designs in a systematic approach. Unlike other domains, in the educational domain there are no studies on AI-design where human values have been systematically approached and structurally integrated, leaving a scientific gap.
This study therefore asks the question:
How can a training in feedback provision, using an AI-driven Avatar as critical trainer, be designed in alignment with staring teachers’ values and concerns?
To answer the question, this study focuses on three research objectives, investigated through three studies: (1) eliciting human values through literature; (2) empirically investigating human concerns, together with the technical design; (3) validating the design through starting teachers’ evaluations. As a validated approach, the Value Sensitive Design Method will be used to elicit values and concerns from practice and theory respectively. At the core of the empirical phase (study 2) is the Q-method that synthesises the theoretical and empirical results, facilitating an iterative mixed-method approach that enables insight into relations, priorities and consensus of users’ values, followed by the technical design and starting teachers’ evaluations.