Why AI Won’t Fix Bad Pedagogy - And What It Can Fix

Every time a new technology appears in education, the same myth resurfaces: “This will fix teaching.”

AI is now the latest target of this misunderstanding. Some vendors promise miraculous outcomes - automated lessons, automated personalised learning, automated everything.

The truth is simpler and more uncomfortable:

"AI won’t fix bad pedagogy. It will only make bad pedagogy faster."

But used well, AI can dramatically strengthen good teaching practices by eliminating the wasted effort that stops teachers from doing their best work.

Let’s separate the hype from the reality.

Why AI Can’t Replace Good Pedagogy

1. AI can’t build relationships - the foundation of learning

Learning is relational. Students don’t engage with content; they engage with people.

AI cannot:

• read the emotional atmosphere of a classroom
• sense when a student is disengaged or anxious
• build trust, belonging, or psychological safety
• tailor motivation strategies to individual personalities

These are core pedagogical acts - and they remain human.

2. AI doesn’t understand the social dynamics of a classroom

AI can suggest activities. It cannot see:

• who needs a confidence boost
• who needs more challenge
• who is masking confusion
• who just had a rough morning at home

Pedagogy is situational. AI lacks situational awareness.

3. AI cannot set meaningful learning goals

AI can generate learning objectives, but it cannot:

• align them to a teacher’s long-term plan
• understand local curriculum nuance
• incorporate student histories or group dynamics
• make value-driven choices about what matters most

Goal-setting is a fundamentally human act that reflects professional judgment.

4. AI cannot make ethical decisions about learners

Teachers continuously make decisions balancing:

• equity
• fairness
• support vs independence
• challenge vs overload
• wellbeing vs performance

AI may assist, but it cannot own these decisions.

What AI Can Fix: The Hidden Workload That Breaks Good Teaching

Although AI cannot repair weak pedagogy, it can eliminate the administrative and cognitive burdens that prevent good pedagogy from happening.

1. AI can remove planning overload

Great lessons require:

• alignment to curriculum
• scaffolding
• differentiation
• sequencing
• assessment integration

Doing this manually takes hours. AI can handle the structural and repetitive parts so teachers focus on pedagogy, not paperwork.

2. AI can support evidence-based practice

AI can:

• suggest research-backed strategies
• recommend spacing, retrieval, and interleaving
• propose formative assessment methods
• generate multiple representations of a concept

This gives teachers more options - without more workload.

3. AI can support reflection and refinement

AI can:

• analyse lesson plans for gaps
• suggest pacing adjustments
• highlight alignment issues
• generate alternative approaches

This strengthens teaching, not replaces it.

4. AI can personalise resources, not relationships

Personalised content ≠ personalised pedagogy.

AI can:

• adjust reading levels
• produce differentiated tasks
• create variants of activities
• scaffold explanations

But applying those resources effectively still requires professional teaching expertise.

What Schools Must Get Right

1. Stop treating AI as a replacement

AI is augmentation, not automation. Schools focusing on replacement are the ones that fail.

2. Train teachers in workflow integration, not just tools

Teachers don’t need technical training - they need workflow strategies:

• When should AI be used?
• When not?
• How does AI fit into lesson cycles?
• How do you maintain pedagogical control?

3. Protect teacher judgment

AI should offer options, not outcomes.

Tools that force decisions instead of supporting them damage trust and harm pedagogy.

4. Use AI to free time for relational work

If AI gives teachers back:

• 10 hours a week
• evenings
• cognitive clarity

…then great pedagogy becomes possible again.

When AI removes the invisible burdens draining teachers, pedagogy improves naturally. Not because AI is “smart” - but because teachers finally have the time, space, and clarity to teach the way they always wanted to.