Explore

New Frontiers

From adaptive learning systems to AI-powered tutoring, education is entering a new era. These aren't distant predictions — they're happening now in classrooms and training rooms worldwide.

Artificial intelligence is no longer a futuristic concept in education — it's a present reality. Generative AI models can create personalized lesson plans, generate assessment questions tailored to individual student levels, and provide instant feedback on written work. These capabilities are fundamentally changing the relationship between educator and technology.

Adaptive learning platforms use machine learning to continuously adjust content difficulty and presentation style based on student performance. Early results show significant improvements in learning outcomes, particularly for students who struggled with one-size-fits-all approaches. The technology isn't replacing teachers — it's giving them superpowers.

The frontier extends beyond content delivery. AI is enabling new forms of assessment, predictive analytics for student success, automated administrative tasks, and even emotional intelligence tools that help educators understand student engagement in real time.

The frontier is not the technology. It is the practice.

It is tempting to measure a frontier by what the models can do. We think that is the least interesting measurement available. Model capability has, for several years now, run well ahead of anything most classrooms are doing with it — the constraint is not what the tool can generate, it is what a learner is actually asked to do with it, and what happens to their understanding as a result.

So we read every frontier claim the same way: what does this change about the learner's process? A tool that produces a finished essay in four seconds has changed the artefact and not the learner. A tool that gives a learner twenty attempts with feedback, where they previously got two, has changed the learning. Both are new. Only one is a frontier worth crossing.

What AI genuinely makes newly possible

One thing on the list is different in kind from the rest: deliberate practice at scale. Mastery has always required many attempts against fast, specific, individual feedback — and that feedback has always been the scarcest resource in education, rationed by how many hours a skilled human has. It is the reason a violinist with a private teacher and a writer with an attentive editor progress differently from everyone else.

This is the first technology that can supply that feedback loop to every learner at once. Not the knowledge — the knowledge was already free, and free knowledge did not produce mastery. The correction. We have written this argument out at length in What AI Finally Makes Possible, and its precondition — that there are still no shortcuts, only more repetitions — in No Shortcuts to Mastery.

Two futures sit inside the same tool

An electric bicycle can be pedalled with assistance, and the rider gets stronger. Or the throttle can be held down, and the rider arrives at exactly the same place having done nothing at all. The machine is identical in both cases; the outcome for the human body is opposite. Intelligence behaves the same way — it is built by effort and it atrophies without it.

This is why augmentation, not replacement is a design position here and not a slogan. The frontier that interests us is the one where AI removes the mechanical work that was never teaching anyone anything, and returns the difficulty that was. See Intelligence Is a Muscle and, for six concrete versions of it, Amplification, Not Substitution.

New formats, not only new tools

The frontier moves in the shape of the learning experience too, not just its instruments. The Smoother methodology treats a learning experience as a transformation the participant undergoes rather than content delivered to them, and the newest thing we have added to it is nano-learning: units small enough to fit in the gaps of an adult life, sequenced so the smallness does not cost coherence.

Underneath sits the Impact Arc — Engage, Enable, Inspire, Empower, Connect — Carlos Miranda Levy's frame for intervening without creating dependency. It is an organisational conviction rather than a validated framework, and we say so plainly wherever we use it. Applied to a frontier technology it asks an uncomfortable question of every new tool: does this build the learner's capability, or does it quietly do the work for them and call that progress?

How we read a frontier claim

Frontier coverage is where fabricated numbers travel fastest, and we are not exempt: we once built an entire infographic on AI-generated figures, checked them, and did not publish it. That discipline — a claim is cited, or it is directional, or it does not run — is documented in The Infographic I Didn't Publish and applies to every page on this site, including this one. It is also worth knowing that every wave of educational technology arrived with the same two prophecies, ruin and salvation, and neither has ever landed: The Recurring Panic.