Constructive alignment with AI: how tooling makes the difference

What is constructive alignment?

Constructive alignment is the instructional model by John Biggs (1996): learning objective, learning activity and assessment are in line. If your objective is “the student can take a patient history”, then the student practises taking a history in the lesson (activity) and the assessment tests that too (question: “take a patient history for this case”, not: “name three parts of a patient history”).

Simple in theory, brutal in practice. Research consistently shows that only 2 to 4 out of 10 lessons are fully aligned. The rest have a break somewhere — an objective that doesn't return in the activity, an assessment that asks something other than what was practised, a summary that uses a different frame than the build-up.

Why common tools break alignment

A typical lesson presentation in PowerPoint has this structure: slide 1 title, slide 2 learning objectives, slides 3-10 explanation, slides 11-13 exercise, slide 14 quiz. The three elements of alignment stand apart from each other. Literally: different slides, different moments, different frame.

What happens across that distance is predictable:

• Slide 2 claims the objective is “taking a patient history”, but slide 5 only explains patient-history theory
• Slide 11 practises with a role-play, but slide 14 assesses with multiple-choice questions about theory
• The summary on slide 13 summarises the theory, not the professional situation the objective refers to

Not a bad lesson. Not an inattentive teacher. Just a tool that doesn't enforce alignment. PowerPoint asks “what goes on this slide?” — never “what was your objective again, and does this assess it?”.

What AI can and cannot solve

It's tempting to think: AI solves this on its own. You type in a topic, AI writes a lesson, done. That is both true and dangerous.

Where AI does help:

• Consistent structure across lessons — every lesson follows the same logic
• Generating assessment questions that match the just-introduced concepts (alignment-by-proximity)
• Building cases in which the objective is practised, not described

Where AI does not help — and even harms:

• If the AI doesn't impose an instructional structure, you get a faster-generated but equally broken lesson
• An “AI slide generator” without flow logic only makes the surface faster, not the content better
• AI can hallucinate — which in an educational context means an objective-activity-assessment triad is structurally sound but factually untrue

The difference lies in whether the AI follows an instructional framework, or merely does a prompt-to-slides mapping. Without a framework: faster but bad. With a framework: structurally aligned.

Alignment-by-design with instructional flows

LectaMe works with three flows. Each framework is a different type of alignment. You choose the flow that fits your objective, and the alignment follows.

The point is not that these three flows are magically aligned. The point is that they put the alignment logic into the format. You can still make a Theory Flow lesson that is poorly aligned (for example by not keeping the checks seriously tied to the concept), but the default is aligned. With an empty PowerPoint slide, the default is nothing.

Concrete example: patient history, MBO-4 nursing

Objective: the student can take a structured patient history from a patient with abdominal pain following the SCEGS model.

In PowerPoint (risk: not aligned): Slide 2 names the objective. Slides 3-8 explain SCEGS theoretically. Slide 9 gives an example question per letter. Slide 10 says “practise together”. Slide 11 is a multiple-choice question about which letter stands for “social”. The assessment demands: the student must take a patient history during a clinical placement — something not practised in the lesson.

In Case Flow (aligned): The lesson opens with a case — “Maria, 34, arrives with abdominal pain at the GP out-of-hours clinic”. The student analyses: what do you ask first? The layer underneath (deepening) opens the SCEGS model when the student gets stuck. Next layer: a second case, different presentation. Assessment is integrated: the analysis questions during the case are the assessment. Objective (taking a history), activity (taking a history in the case) and assessment (evaluating that history) coincide.

Same topic. Same theory. Different alignment. And — not unimportant — a completely different student experience.

Implementation tips

Three things that help keep alignment consistent, with or without LectaMe:

1. 1Formulate objectives as verb-action, not as noun-knowledge. “The student knows the SCEGS model” is not assessable without translation. “The student applies SCEGS in a patient history” is directly assessable.
2. 2Choose your flow by your objective, not your topic. “Anatomy of the heart” sounds like knowledge transfer (Theory Flow), but if the objective is “the student recognises abnormalities”, Case Flow is better.
3. 3Check after generation: does my objective return in both the activity and the assessment? This is a 10-second check. If the answer is “not directly”, adjust the assessment or activity — or switch the flow.

When you should not rely on AI alignment

Alignment-by-design via AI flows works well for regular lessons — weekly knowledge transfer, cases, assessments. It works less well for:

• Large summer programme redesigns where alignment is tested curriculum-wide — there human expertise leads
• High-stakes summative exams — never let AI generate final assessments on its own
• Inquiry-driven learning where objectives evolve during the lesson — there alignment is a moving target

For those cases AI support is at most a first draft. For 80% of weekly lessons, alignment-by-flow is a structural improvement.

Sources

• Constructive alignment — John Biggs (original concept) — the instructional foundation that aligns objectives, activities and assessment.
• Times Higher Education — international articles on applying constructive alignment in higher education.
• Vernieuwenderwijs.nl — Dutch instructional blogs about education and evidence-informed practice.

Further reading