AI is reshaping education: the chances are, in the last week, you’ve seen or heard colleagues talking about both the opportunities and the threats it poses to traditional teaching and learning practice.
Nowhere is this contrast more stark than in the context of accessibility. On the one hand, new technologies mean, for instance, those pupils who have struggled for years to decipher handwriting in their books and on whiteboards have a viable alternative.
On the other hand, a widespread switch to devices risks leaving behind all those pupils living, or attending school, in socioeconomically deprived areas.
In this blog we unpack the implications of AI when it comes to accessibility, and consider whether it’s possible to enjoy the best of both worlds.
What does ‘accessible’ really mean?
When we talk about accessibility, particularly in the context of technology, we’re usually concerned with the number of people for whom a product or service is usable.
Designers increasingly take this into account in their font and layout choices when designing apps or websites, and with the addition of features like image descriptions or closed captions on videos.
In an education context, teachers will be familiar with the process of providing enlarged-text papers or coloured backgrounds to support pupils with visual impairments. Where this isn’t possible — as in the case, for instance, of well-worn copies of set texts or textbooks — screen readers can plug the gap.
There is another side to accessibility that exists beyond the classroom, however: the number of pupils for whom a product or service is usable at home.
Lack of internet, or lack of a device like a phone or laptop, can make virtually any online solution inaccessible to a large number of pupils, particularly in areas of high deprivation. The same can be said for schools, where lack of reliable infrastructure or working devices can mean not all classes or pupils benefit from a resource.
Evaluating accessibility
When we are evaluating how ‘accessible’ an educational resource is, we can therefore ask two simple questions:
For what percentage of our cohort is this resource usable in school?
For what percentage of our cohort is this resource usable outside school?
Large disparities between the answers to question 1 and 2, or a low percentage for either question, risks widening the gap between groups of pupils — schools should proceed with such resources with caution.
To screen or not to screen?
Screens bring with them a host of accessibility benefits, including:
Read aloud function for those with visual impairments or low literacy levels
Ability to zoom in or resize text and images
Translation functions to support EAL pupils and their families
Image descriptions
Writing-to-text software to turn handwritten notes into typed text
Then there’s the benefits from a pedagogical perspective:
Portability of huge volumes of information that couldn’t be contained in a single textbook (at least one that would fit in a school bag!)
Instant feedback
Interactivity
Ease of storing, sharing and sending information or ideas
Increasingly sophisticated chatbots and LLMs to support planning and feedback
Some exam boards in the UK are already sharing plans to move to online exams, perhaps the strongest signal of all that education is moving towards screens and away from paper.
For pupils used to struggling to get down all the complexities of their thoughts with a pen and paper, or battling to read yet another page of too-small print, this news will be a welcome relief.
Some schools are even experimenting with the use of ChatBots as virtual tutors, or AI image generators to offer pupils a novel way of interacting with the imagery in literary texts. Such strategies capitalise on the responsive nature of AI to provide, in turn, a form of responsive teaching.
But screens are not the panacea they might seem when it comes to improving pupil outcomes.
A study by John Jerrim analysed the results of over 3000 pupils taking the 2015 PISA test, with pupils divided into two groups: those taking the test on paper, and those taking it on a computer. The analysis found pupils taking the test on paper outperformed the computer-based group by 20 scaled score points. Such a significant difference has serious implications when considering a move to online assessments — but why might it exist in the first place?
Peps McCrea has written about ‘screen inferiority’ and possible reasons why, particularly where longer texts are concerned, pupils may demonstrate lower levels of understanding on screen compared to on paper.
This can partly be attributed to the distractions screens pose: even with notifications turned off and ‘Do not disturb’ turned on, our brains quickly learn to look for more than just text. The need to scroll also brings with it the need to ‘locate’ a text on the page, something that isn’t necessary on paper.
To prepare pupils adequately for online assessment, they would need to spend a substantial proportion of their time using devices as part of their preparation — the potential impact of this on learning cannot be dismissed.
What's more, at least a base level of training is needed for both teachers and pupils before AI-driven platforms can be used safely and effectively. AI is not infallible; if it is relied on as an additional teaching resource for those pupils needing support, proper checks need to be in place before, during and after usage to protect against misconceptions or exposure to unsuitable material.
The best of both worlds
At Stylus we’re keeping paper at the heart of the process: this means even pupils with poor access to devices at home or in school can enjoy detailed additional feedback through our LearnCycle service.
Though we use AI at different points during a LearnCycle, teachers and pupils have no direct interaction with this part of our process, and we moderate our outputs at every stage to ensure our feedback is of the highest quality.
That doesn’t mean, however, pupils can’t benefit from the advancements in accessibility that screens can offer. In the future, these could include:
Automatic resizing or double spacing of question papers for those pupils requiring it
Automatic addition of coloured backgrounds for those pupils requiring them
Online answer submission for pupils whose usual way of working is on a laptop
Translation options for pupils or parents who don’t speak English as as first language
We want to hear from you: what else do you, and your pupils, need for better accessibility? If there is something you think we’ve missed, get in touch: hello@stylus.education