Aktivite 5
This is the space to describe the service and explain how customers or clients can benefit from it. It’s an opportunity to add a short description that includes relevant details, like pricing, duration, location and how to book the service.
Aktivite 6
This is the space to describe the service and explain how customers or clients can benefit from it. It’s an opportunity to add a short description that includes relevant details, like pricing, duration, location and how to book the service.
Aktivite 7
This is the space to describe the service and explain how customers or clients can benefit from it. It’s an opportunity to add a short description that includes relevant details, like pricing, duration, location and how to book the service.
Aktivite 8
This is the space to describe the service and explain how customers or clients can benefit from it. It’s an opportunity to add a short description that includes relevant details, like pricing, duration, location and how to book the service.
Low-stakes checks first
Quick retrieval quizzes are a good starting point. An AI assistant can draft ten multiple-choice questions on a topic in seconds, including plausible wrong answers. The teacher's job is to check accuracy, remove ambiguous items and make sure the questions cover more than recall. Bloom's taxonomy is a helpful reference here: ask for questions that require pupils to explain, compare or apply, not just remember.
Automatic marking of closed questions is long established in platforms such as Google Forms and Moodle quizzes. It is reliable because the answer key is set by a person.
Keep a short bank of checked questions so they can be reused and improved each year.
Feedback on open answers
Tools that comment on essays or short written answers can give pupils something to act on quickly. Their comments can also be generic, miss the point of a strong but unusual answer, or treat fluent writing as correct reasoning. Treat AI feedback as a first reader, not a final one, and tell pupils clearly when a comment came from a tool.
For any assessment that counts towards a grade, a teacher should read the work and decide the mark. Sample moderation within a department helps check that any AI-assisted process is not drifting.
Fairness needs particular attention. Automated feedback may respond differently to pupils writing in a second language, or to dialect and spelling patterns it has seen less often. Checking a sample of feedback across different groups of pupils, and asking pupils whether the comments made sense to them, helps catch this early. Schools should also be clear about when pupils may use AI in coursework, so that the evidence teachers assess genuinely shows what each pupil can do.
A self-assessment activity with AI feedback
This works with Year 10 (ages 14 to 15) in a 40-minute English lesson. Materials: each pupil's draft paragraph, a printed success criteria sheet, and access to a school-approved AI feedback tool.
Pupils mark their own paragraph against the success criteria first (10 minutes).
They then ask the tool for two strengths and one improvement linked to the same criteria (10 minutes).
They compare the two sets of comments and highlight where they agree or differ (10 minutes).
They redraft one sentence and write a short note on which feedback they chose to follow and why (10 minutes).
Key points
Use AI for drafting quiz items, then check every question yourself.
Keep grade-bearing marks with teachers and moderate samples.
Label AI-generated feedback so pupils know its source.
Ask pupils to compare AI feedback with their own judgement.
Photos: Yustinus Tjiuwanda, Rafay Ansari / Unsplash
AI in Assessment and Evaluation
Assessment takes a large share of teacher time, and AI tools promise to speed up question writing, marking and feedback. Some of that promise holds up, particularly for low-stakes checks. For work that affects grades, reports or pupil pathways, teachers need to stay in control of every judgement.
Edutech
/
/
AI in Assessment and Evaluation
