Can AI help to address the literacy skills gap?
Stephen Park, MD at Lexplore, discusses reading skills assessments and how AI could be used to improve and streamline processes.
Early disadvantage can be devastating in the classroom. We know, for example, that bright primary school children in receipt of free school meals struggle to keep up with their better-off peers. According to new analysis from the DfE, they are also less likely to be in lasting employment at the age of 27.
Additionally, we know that the disparities begin to show right from the start of education. More than a quarter of children finish Reception without the communication and reading skills they need to flourish. It’s a situation highlighted by Damien Hinds’ recent pledge of £20 million to address the attainment gap. As the education secretary noted in his speech, “When you are behind from the start you rarely catch up.”
At the coalface, many schools are already working tirelessly to boost literacy skills. After all, it’s widely recognised that this is the starting point for access to the wider academic curriculum. Yet the complexity of how we learn to read is sometimes underrated and arguably one already stretched teachers are underprepared for. Could it be time to introduce AI as a new approach to this age-old problem?
A learning journey
When it comes to the acquisition of reading skills, there remains a tendency to assume that, with the right encouragement and opportunities, children will pick it up in the same way that they pick up language skills.
But there is a huge difference. We are hardwired to pick up language. Reading, on the other hand, is a code that we need to learn how to decipher. And it’s not straightforward. It might surprise you to discover, for example, that in English, the sound ‘sh’ can be made 13 different ways.
So children need to learn when the rules apply, but also when they don’t. Those with a good visual memory will increasingly remember and recognise the sequence of letters in words for the next time it crops up, but for some children, this becomes an impossible task. And if they manage to develop coping strategies which temporarily hide the fact they are struggling, the problem becomes harder for educators to spot and support.
Focus on skills
From our own experience of using AI in schools to uncover reading issues, we have found it helps circumnavigate many of these issues. There is an established and well-documented correlation between reading processes and its various manifestations in eye movement. Tracking how a child’s eye moves measures how effectively the cognitive processes related to reading work together. It can look at when, where and how the eye moves in relation to the words, sentences and text the child is reading.
The same algorithms can also be used to detect dyslexia, the result of the brain being unable to decode text, without the need for lengthy waits to be assessed.
Systematic and objective identification of any issues could greatly impact how we support reading attainment. Currently, past Key Stage 2, curriculum demands mean teaching training cannot fully cover how to teach an older child to read. Nor can all teachers – or teaching assistants who are often the ones to help those struggling – be expected to be dyslexia specialists.
Testing literacy is also extremely complicated, as it encompasses a fine interplay of different cognitive and linguistic processes when it comes to fluency, speed and comprehension. At the moment, screening tests such as Year 1 phonics are riddled with pitfalls. They do not necessarily assess the full range of phonics knowledge that the national curriculum requires; teaching single sounds doesn’t equip children with comprehension skills.
AI is the type of technology that can really make a difference. It does not even require children to have fully developed communication skills, which means that a child starting school with few verbal competencies can still be accurately assessed. And it can also do it much quicker than any other reading assessment because it uses data from, in our case, 30 years of eye-tracking data to very accurately spot any issues. In fact, it can take as little as two minutes to assess each child.
At a time when the perils of social immobility are so severe, it’s imperative that we do all we can to ensure no child is left behind. Their best chance is early intervention, which relies on spotting literacy issues at the earliest possible opportunity. This is the only way that time and resources can be used effectively to help students who need the most support in their reading development. And AI encourages a level playing field, ensuring that any support has the greatest chance of success.