A chance for success: a new vision for inclusion in education

We present to the reader the reflections of Paddy McGRATH, a world expert in the field of global educational technology. This material is based on his speech at the AI for Education forum in Almaty, which has become one of the most prominent events in the recent discourse on the future of inclusive education.

How can we ensure that high-level innovations deliver meaningful local results within the modern classroom? What is inclusion in its fullest sense? Which students need AI assistance, and does everyone have the opportunity to receive it to achieve academic success? These are the questions I seek to answer in my research, which I conduct in various countries around the world.

Everway, the company where I work, has provided inclusive education tools to more than 250 million students worldwide. Working at such a scale, we must ask ourselves: how do we actually turn global innovation into real local results?

It all begins with recognizing our reality. Technology is evolving at a breakneck speed; AI is no longer a concept of the future, but an inevitable presence in our classrooms, and with this, expectations are rising. Yet, despite all this technology, gaps persist – between rural and urban areas, between different socio-economic backgrounds, and between students of varying abilities.

The biggest challenge we face is not a lack of software, but a lack of time. Every teacher knows that one-on-one is the gold standard of support, but in a crowded classroom, it is the hardest thing to provide.

To solve this problem, we need to shift our perspective on inclusion.

I like to use the iceberg analogy. Usually, we focus on the visible tip – students with diagnosed physical or cognitive impairments. But the vast majority lie "below the waterline." These are the students who are simply struggling to keep up, or children with a language barrier, or those who have a mild form of dyslexia but have not yet been "labeled." AI gives us a chance to support the entire iceberg.

Consider the architecture of inclusion. I often show a photograph of the steps at City Hall in Vancouver. Once upon a time, instead of just building a massive staircase and then "bolting on" a clumsy lift for wheelchair users, they designed the steps so that the ramps were integrated right into the stone. It is beautiful, functional, and, most importantly, it is for everyone. It is used by many: parents with strollers, travelers with heavy suitcases, and the elderly. This is exactly how we should treat AI in education. It should not be some "special" tool hidden in a corner; it must be part of the standard design of the learning process.

Let us look at specific examples of how this works. We now have tools that provide natural AI voices for text-to-speech, real-time translation, and the simplification of complex vocabulary. For a student learning in a language that is not their native tongue, or for one struggling with literacy, these are not just "cool features" – they are a lifeline. But here is where the magic happens:

when these tools are available to every student, the very notion of a "stigma" disappears.

When a top student uses a simplification tool to understand a complex text, it normalizes the use of that same tool for a struggling student. We give them choice and a voice. We empower children with independence – the opportunity to use a tool exactly when they need it, without raising their hand and saying, "I am different, I need help." This opens the way to solving Bloom's famous "2 Sigma Problem," which proves that one-on-one tutoring yields a colossal leap in performance – and AI is capable of providing such an adaptive digital tutor to everyone.

Scaling this approach is difficult because, frankly, many strategic plans claim to be inclusive, but if you dig deeper, they are not. Too often, inclusion is viewed as something extra, an "add-on" – something we think about only after we have designed the curriculum for the so-called "mainstream." Reforms are often written backward, attempting to fit children with special needs into a ready-made system, rather than changing the very concept of mass education.

I try to avoid using this word – "mainstream" – because it implies the existence of some "normal" group and everyone else. We need to move away from a model where everything relies on a few passionate "enthusiasts" in a school. We need a system-wide standard.

The lesson I have learned from working in different countries is that only simplicity scales. If a tool is too complex, a stressed and overworked teacher will not use it. But if we embed simple and powerful AI support features – such as high-quality dictation or automatic captioning – into the daily workflow, it becomes good practice, rather than just "practice for people with disabilities."

I want you to remember this: scale the outcome, not the technology. When you look at an AI tool, do not just ask what it can do; ask what barrier it removes. What is necessary for some is useful for all. If we design for the "edges," we ultimately support the "middle" as well, and that is how we truly guarantee every student the opportunity to be successful.