RETHINKING LEARNING SUPPORT
Education in the 21st Century & a New Model for Learning Support
Education in the 21st Century
Education today exists within a world defined by complexity, rapid change, and constant cognitive load. Young people navigate environments saturated with information, competing demands on attention, and increasingly diverse pathways to knowledge. In this context, the traditional model of teaching — built around standardisation, linear progression, and remediation of perceived deficits — is no longer fit for purpose.
The 21st century requires learners with developed metacognitive and epistemic awareness skills, who can think flexibly and critically, regulate and motivate themselves, make informed decisions, navigate complexity and fast change, and adapt to new challenges. These are not “extra” skills; they are the foundation of participation in modern life.
Content is abundant. Capability is scarce.
The shift from performance to capability
For decades, schools have been structured around performance indicators — grades, levels, targets, and assessments. But performance does not automatically translate into capability. A student may achieve high marks while still lacking the ability to plan, monitor, or evaluate their own learning.
In the 21st century, capability matters more than compliance. Learners need:
• the ability to manage cognitive load
• strategies to navigate complexity
• confidence to make decisions
• metacognitive awareness
• autonomy to act without constant adult mediation
These are the skills that enable long‑term independence — and they cannot be developed through remediation‑first approaches.
Why neurodiversity reframes the conversation
Neurodiversity is not a marginal consideration; it is a central feature of modern classrooms. Cognitive variability is the norm, not the exception. When systems are designed around a narrow definition of “normal,” learners who process, organise, or communicate differently are positioned as problems to be fixed. But when systems are designed around variability, not uniformity, every learner benefits.
The 21st century demands environments that remove barriers, not identities.
Remediation focuses on correcting weaknesses in literacy, numeracy, working memory, processing speed, or executive functioning.
The Problem with Remediation‑First Models
What Traditional Learning Support Gets Wrong
-
Over‑focus on remediation
-
Narrow definitions of “need”
-
Fragmented interventions
-
Low expectations masked as support
-
Systems that create dependency
A remediation‑first ecosystem limits capability rather than expanding it.
For many neurodiverse learners, this approach:
-
delays access to the curriculum
-
reinforces a sense of inadequacy
-
consumes time and cognitive energy
-
rarely generalises to real‑world independence
-
positions the learner as someone who must be “fixed”
-
is support-dependence provoking
Why traditional models no longer meet the demands of modern learning
Traditional SEN Support and LSA-Dependence
In England, SEN provision has long relied on Learning Support Assistants (LSAs) as the primary mechanism for supporting students with learning difficulties. While LSAs are essential for some learners with complex and significant needs, the structural model creates dependence rather than autonomy.
This 'velcro adult' model can:
- reduce access to high-quality teaching
- limit opportunities to develop self-regulation
- create overscaffolding that substitutes for thinking
- reinforce the belief 'I can only do this with an adult next to me'
- stigmatise learners
- delay and/or prevent the development of metacognitive skills
- prevent the development of transferable strategies
The velcro-adult model creates dependence, not capability.
Compensation:
A More Empowering Framework
Compensation removes barriers rather than removing difference. It gives learners the tools they need to participate fully now, while building long‑term capability.
What compensation looks like:
Planning for Variability (UDL Principles)
• Multimodal access to content
• Personalised scaffolds for comprehension
• Tools that reduce cognitive load without reducing challenge
• Flexible pathways for engagement and expression
Compensation as Standard Practice
• Text-to-speech and speech-to-text
• Multimodal output options
• Strategic use of AI-learning tools
• Assistive technologies embedded into everyday learning
Explicit Teaching and Coaching
• Approaches to learning & AI use taught explicitly and revisited over time
• Mindset coaching that strengthens resilience and agency
• Explicit teaching of self-advocacy skills
• Guided practice with gradual release
Compensation is not a shortcut — it is the foundation of capability. It feeds directly into the SRL cycle by expanding strategy repertoire, supporting effort regulation, and enabling accurate self‑evaluation.
Universal Design for Learning (UDL)
Designing for Variability
Compensation only works sustainably when it sits within a universally designed environment. Universal Design for Learning (UDL) provides the system‑level architecture for this shift: it removes unnecessary barriers, offers multiple pathways for engagement and understanding, and recognises variability as the norm. When UDL principles shape the learning environment, compensation becomes a natural extension of design rather than an exception or an intervention. This alignment ensures that autonomy is built into the system, not added on top of it.
AI as a Catalyst for Autonomy
AI has the potential to strengthen autonomy when it is aligned with the SRL cycle. Used well, AI becomes a thinking partner that supports strategy development, decision‑making, cognitive empowerment and reflective growth.
How AI strengthens the SRL cycle
Strategy Development
AI exposes learners to diverse approaches, models expert thinking, creates resources and expands their strategic repertoire.
Strategy Selection & Implementation
AI prompts learners to choose strategies, sequence steps, and apply them effectively — without taking over.
Effort Allocation
AI helps learners estimate time, manage cognitive load, and regulate emotional effort.
Self‑Evaluation
AI supports accurate self‑assessment by helping learners compare intentions with outcomes.
Reflection & Adaptation
AI guides structured reflection and helps learners refine their personal learning profile.
AI strengthens autonomy when it supports the cycle — not when it replaces it.
A Self‑Regulated Learning (SRL) Model
for AI‑Enabled Autonomy
Building capability through strategy, awareness, and adaptive thinking
Self-Regulated Learning (SRL) is a cyclical, developmental process through which learners build the internal systems that allow them to plan, act, adapt, and grow independently - and lead their own learning journeys. It is not a single skill — it is a repertoire of cognitive, metacognitive, motivational, and behavioural processes that strengthen over time. The five‑stage SRL cycle provides a clear framework for understanding how autonomy develops — and how AI can support it.
The SRL cycle — Strategy Repertoire → Strategy Use → Effort Allocation → Self‑Evaluation → Reflection & Adaptation — is particularly powerful for neurodiverse learners because it makes thinking visible, transferable, and improvable. When aligned with SRL, AI becomes a cognitive partner that strengthens autonomy rather than replacing effort.
Why SRL Matters for AI‑Enabled Autonomy
When AI is aligned with the SRL cycle, it strengthens the learner’s internal systems:
- Strategy repertoire → expands possibility
- Strategy selection → builds decision‑making
- Effort allocation → builds regulation
- Self‑evaluation → builds accuracy
- Reflection → builds adaptability
Risks: When AI Reinforces Remediation
Without clear standards and explicit teaching and guidance, AI may drift toward deficit‑based practices:
- automated marking that penalises neurodivergent communication
- behaviour-tracking tools that misinterpret ADHD traits
- skill-drill apps that correct deficits rather than enable access
- tools that overscaffold and reduce opportunities for metacognition
- systems that collect sensitive cognitive or behavioural data without transparency
AI must not become a new mechanism for normalising learners.
Compensation is not about making tasks easier — it is about making them possible.
What Good Looks Like
AI should be evaluated through the lens of autonomy, not automation.
High‑quality AI tools:
- reduce unnecessary cognitive load
- increase access to content
- support planning without taking control
- strengthen metacognition
- respect neurodiverse communication and processing styles
- avoid pathologising difference
The future of learning support is not remediation — it is autonomy.
AI offers a once‑in‑a‑generation opportunity to redesign learning support with autonomy at its core. For neurodiverse learners, this shift is essential. With the right pedagogical and regulatory frameworks, AI can expand access, strengthen independence, and enable learners to participate fully and confidently in their education.