SILT
00%
SILT
00%
Skip to content
SILT
Our OriginOur research

Research

How do we ensure human learning evolution - not just model performance - in a rapidly evolving AI-embedded world?

More time with AI is not necessarily more learning. Our research agenda asks what helps people understand, remember, explain, and act independently, while exploring cognitive health, wellbeing, and behavioural health across lifelong learning.

These are research directions and hypotheses, not claims of demonstrated outcomes.

Follow the inquiry

The critical gapOur research directionAI, adolescence & social connectionPersonal feedback. Classroom scale.The effort that makes learning happenMemory, privacy & classroom governanceCognitive health, wellbeing & behavioural healthWhat becomes a stronger measure of learning?From questions to practiceCurrent case study

The problem we are investigating

Lifelong learning is a continuous path of discovery that shapes our minds and hearts from childhood to old age.

While the need for lifelong learning is universal, our educational and corporate systems are still largely built for a one-size-fits-all model, leaving diverse learning needs unsupported at scale

Learners do not arrive with the same conclusion about their prior knowledge, pace, confidence, misconceptions, cognitive strengths, or ways of making sense of new ideas. Yet many educational systems still depend on shared instruction, standardized assessment, and limited opportunities for continuous individual feedback.

This is not simply an educator problem. It is a problem of scale.

Teachers can adapt when they understand what a learner is struggling with, but continuously identifying where every learner is confused, what they have retained, and what kind of support they need next becomes increasingly difficult across large classrooms.

At the same time, access to generative AI does not automatically produce better learning. More answers, more interaction, or more time spent with AI can just as easily reduce productive effort, encourage dependence, or obscure what a learner actually understands.

SILT is investigating whether AI can help close this gap by becoming a personalized learning-feedback layer around existing teaching.

Alongside this classroom inquiry, we are expanding our exploration to cognitive health, wellbeing, and behavioural health across lifelong learning. We want to understand the effects of technology in the varied settings where people learn, work, and participate in everyday life.

Our central question

How can technology and evolving AI support learning, cognitive health, wellbeing, and behavioural health across the lifespan?

These are research directions and hypotheses, not claims of demonstrated outcomes.

The direction of travel

Support learning without replacing the human work through which learning develops.

We are interested in how AI can support learning without replacing the human work through which learning develops.

That means studying systems that do more than generate explanations or answer questions. We are exploring whether AI can interpret evidence from learner interaction - what someone can retrieve, explain, apply, where misconceptions persist, and how they respond to previous feedback and use that evidence to shape what happens next.

The goal is not to create a different curriculum for every learner.

It is to investigate whether a shared learning environment can become more responsive to the individual moving through it.

Our wider exploration asks how systems can support cognitive health, wellbeing, and behavioural health across life stages and learning needs. We aim to observe their effects over time and use what we learn to improve their design.

Connected lines of inquiry

Questions that keep the work honest.

AI, adolescence & social connection

How can personalized learning support strengthen confidence and agency without replacing human relationships?

Adolescence is a period in which learners are developing confidence, independence, communication habits, and beliefs about their own ability. As AI becomes more present in learning, we are interested not only in what it helps young people accomplish, but in how the relationship itself is designed. Personalization may make support more available, but it can also create new questions around dependence, trust, social withdrawal, and when AI should deliberately return a learner to another person.

Questions we are exploring +
  • When should AI answer, question, pause, or encourage a conversation with another person?
  • Can feedback strengthen confidence without encouraging dependence on the system?
  • How should we assess communication, learner agency, and social participation alongside academic performance?
  • When should AI step away so that teachers, peers, or families remain part of the learning process?
  • Can personalized support increase a learner’s willingness to participate, ask questions, and seek human help?

Personal feedback. Classroom scale.

How can each learner receive useful feedback while their teacher gains capacity?

The classroom creates a fundamental tension: learners need individualized feedback, while educators have limited time and attention. We study AI as a supportive learning-feedback layer within existing classrooms, with educators able to review, challenge, and redirect its observations. Rather than asking AI to replace teacher judgment, we are interested in whether it can help make learner understanding more visible.

Questions we are exploring +
  • What should adapt: feedback, questioning, pace, practice, or the form of explanation?
  • How can teacher feedback change the support an AI system provides next?
  • Can specialized agents coordinate practice, assessment, planning, and feedback more usefully than a single assistant?
  • How do we measure the effect on educator workload as well as learner understanding?
  • Can assessment become a continuous feedback process rather than a final score?
  • What learner information is genuinely useful to an educator, and what becomes unnecessary noise?
  • Can AI surface persistent misconceptions or patterns of difficulty without requiring teachers to review every student interaction?

The effort that makes learning happen

How do we use AI without removing the thinking through which understanding develops?

Convenience can quietly remove the cognitive work that makes learning durable. We explore active recall, self-explanation, appropriate challenge, focused engagement, and productive struggle as design principles for AI-supported learning. The question is not simply whether AI can make learning easier. It is whether AI can make the learner increasingly more capable of working without it.

Questions we are exploring +
  • What changes when a learner attempts an explanation before receiving an answer?
  • Can adaptive questions distinguish genuine understanding from familiarity?
  • Which interaction patterns improve later retrieval and independent problem-solving?
  • When should AI stop helping so that productive effort can continue?
  • Can feedback adapt difficulty without eliminating useful struggle?
  • How should an AI system know when a learner is ready to move from support to independence?

Memory, privacy & classroom governance

How much should an adaptive learning system know about a learner?

Personalization benefits from context. Learners deserve boundaries around what is collected, inferred, remembered, and shared. We investigate privacy not as a separate compliance layer, but as part of learning design itself. A system may need enough information to recognize recurring gaps or progress, but it should not accumulate information simply because it can.

Questions we are exploring +
  • What is the minimum information needed for useful personalization?
  • Can learning-state summaries replace unnecessary conversational history?
  • Which information belongs with learners, families, teachers, or institutions?
  • Where could local processing, limited retention, and governed human review reduce risk?
  • Which learner inferences should AI systems be prohibited from making?
  • How should students and educators understand what the system remembers and why?
  • When should historical learner data stop influencing future recommendations?

Cognitive health, wellbeing & behavioural health

How can technology and evolving AI contribute positively to how people think, feel, and live across a lifetime of learning?

We are opening a research exploration into how digital technologies and rapidly evolving AI influence cognitive health, wellbeing, and behavioural health. We are interested in attention, memory, self-regulation, everyday habits, and social participation across school, work, community life, and later life. Our aim is to design, observe, and refine systems around different learning contexts and needs, examining potential benefits alongside unintended effects over time.

Questions we are exploring +
  • How do different patterns of AI use affect attention, memory, and independent thinking over time?
  • How might learning technologies support wellbeing, self-regulation, and constructive everyday habits?
  • How do effects differ across life stages, abilities, cultures, and learning contexts?
  • Which design choices support meaningful activity and human connection beyond the screen?
  • How should learners and relevant cognitive, behavioural, and health specialists shape measures, interpretation, and research boundaries?
  • What evidence would show a positive contribution, no meaningful change, or harm, and how should each finding change the design?

Human capability is the measure

What becomes stronger in the learner in the AI age?

More time with AI is not necessarily more learning. Across these research areas, our broader question is whether intentionally designed AI-supported learning environments can strengthen the learner rather than simply increase access to assistance.

Understanding

Can the learner explain why?

Retention

Can they retrieve it later?

Transfer

Can they use what they learned in an unfamiliar situation?

Reasoning

Can they work through problems rather than wait for answers?

Metacognition

Do they know what they understand, what they do not, and when they need help?

Agency

Can they question, attempt, explore, and make decisions independently?

Intellectual range

Can personalization help learners develop new ways of reasoning, explaining, and solving problems?

Connection

Does technology strengthen rather than displace relationships with teachers and peers?

Cognitive health

How do attention, memory, and independent thinking change with different patterns of technology use over time?

Wellbeing & behavioural health

Does the design support wellbeing, self-regulation, constructive habits, and participation in everyday life?

These are research directions and hypotheses, not claims of demonstrated outcomes.

The work returns to practice

From questions to practice.

SILT’s research is intended to inform the design of learning systems, not remain separate from them.

Through SILT Labs, we can translate research questions into carefully bounded prototypes, pilots, and learning interactions, review how systems behave across educational and lifelong learning contexts, and use that evidence to improve both the systems we build and our understanding of their effects on learning, cognitive health, wellbeing, and behavioural health.

Research asks what should change. Practice helps us calibrate it. Evidence tells us what deserves to continue.

Continue exploring

SILT Labs↗

Research. Practice. Human possibility.

Back to top ↑

The next chapter
needs human insight.

Bring a classroom perspective, a research question, or an idea for collaboration.

Contact us ↗Subscribe to Our Substack ↗
OriginAbout SILTResearchCase studiesLabsPrinciples & safetyPeopleLearning evolution
SILT

Systems for Integrated Learning Technologies

An interdisciplinary research team in learning science and engineering.