
AI Feedback: Formative Assessment Meets the Bias Test
Instant feedback, self-regulation support: AI transforms formative assessment — provided we manage perception biases and autonomy.
Research notes, data sovereignty and the making of sovereign Swiss agentic AI.

Instant feedback, self-regulation support: AI transforms formative assessment — provided we manage perception biases and autonomy.

From non-player characters that converse freely to immersive worlds, generative AI opens new grounds for practising skills.

How to ground an AI tutor in the right content and give it a memory of the learner, without sinking into hallucinations.

Learning data makes it possible to spot struggling students early. Powerful, but undermined by bias and the question of consent.

AI promises to give teachers time back. Where does real adoption stand, for which uses, and under what conditions?

AI promises more inclusive education for learners with disabilities and multilingual audiences, but equity is not automatic.

Educational AI touches minors and learning trajectories. An overview of the rules and governance principles for responsible use.

Agentic AI designs courses, coaches tutors and runs dialogues. A tour of the real promises and the grey areas.

Fallible detectors, unfair accusations, bias against non-native writers: research argues for integrity built on education, not surveillance.

Essay grading, short-answer scoring, 'LLM-as-judge': a status report on a promising automation that still needs humans in the loop.

Tailored study plans, adjusted difficulty, targeted remediation: AI revives adaptive learning — with new pitfalls to know.

Two years after ChatGPT burst into classrooms, research paints a mixed picture of generative AI in learning.