Digital accessibility and AI: What impact will AI have on digital accessibility?

The research team at System Concepts understand what a true privilege it is to be able to speak directly to users about their experiences of using digital products, with the aim of improving this experience. It is particularly exciting to speak to users during a pivotal moment in the evolution of technology as we know it. There is no denying that we are heading towards a world where AI is becoming more and more commonplace, and it is already possible to imagine a future where life without AI will feel archaic! We remain hopeful that humans will be able to create a harmonious future alongside AI, where everyone, including those with accessibility needs, can feel empowered to achieve their goals.

With this in mind, we have summarised several instances where AI is already starting to impact digital accessibility that we have come across.

Descriptions of visual content to aid those who cannot see

One of the great ways that AI can be used to enhance accessibility is when blind or vision impaired individuals can use it to get descriptions of images or any visual information (like graphs or infographics). This can be achieved by a variety of different AI software which is often integrated into various devices, browsers or apps. But perhaps the most commonly known ones are ‘Be my AI’ (via Be My Eyes which is an app many blind users use for describing almost anything) and ‘Seeing AI’ which is integrated into Microsoft. Using these AI tools can really help blind and vision impaired users to understand what exactly is on screen especially if elements/images aren’t coded correctly to make effective use of their screen readers – therefore creating an overall more accessible experience.

However, we must not use this as an excuse to not give descriptions for informative images – AI is going from strength to strength but it can misidentify numbers in a graph or the contextual intention of images. So thoughtful human-led image descriptions remain the default for genuine accessibility.

Summarised content to reduce cognitive load and aid understanding

AI can now generate concise summaries that minimise the need to manually read and digest vast amounts of content. For some neurodivergent individuals, processing extensive pieces of information can be exhausting, time consuming and frustrating. AI can help by summarising and simplifying text (e.g. articles, documents, emails etc.) to make it more digestible. A great example of this is Google’s ‘AI Overview’, which many of us have most likely come across by now. This provides a high-level consensus of the results to your specific search, available at a glance. In addition to this, the way in which the text is visually broken up by AI is also useful. Often AI will structure these summaries with clear headings and snappy bullet points making it far easier to digest for everyone, in particular for those neurodivergent individuals who struggle with information processing.

Auto captions to aid those who cannot hear

AI has truly revolutionised the experiences of Deaf and hard-of-hearing individuals by providing real-time automated captioning for both live and pre-recorded content. For example, Automatic Speech Recognition (ASR) can be used to generate live captions automatically during video calls (e.g. Zoom), making it possible for these individuals to understand what is being said. Captions can also be added to pre-recorded content, for example video streamers such as YouTube have provided for a while now the option to add automatic captions to videos, and modern AI has significantly improved the accuracy and timing of these captions. This automation makes auditory content and calls so much more accessible for those that may struggle to hear. However, we must remember that for the Deaf community, sign language remains their first language – the one that feels the most natural and easiest to understand. So, while AI captions are an incredible tool, we must continue to provide sign language alternatives to ensure we are truly respecting the needs of all users.

‘Intelligent’ predictive auto suggestions to aid meaningful communication

The use of predictive text and auto-suggestions is a staple of how most of us type. However, AI tools have elevated this experience, by providing context-aware suggestions that predict entire phrases based on the user’s intent – rather than simply finishing a word once it is partially typed. This sort of personalised support is helpful to all, but is life changing to those managing fatigue or physical pain due to dexterity or upper limb impairments as well as reducing the mental effort required for spelling which is challenging for neurodivergent users, like those with Dyslexia. As a result, AI minimises any frustrations and additional tiresome effort needed for both these types of users.

What the future may hold:

It is a genuinely exciting time to be working in the digital accessibility field. With ‘Vibe Coding’, and enhanced AI support for all of us, we are entering an era where ‘Hybrid Automation’ allows us to use AI to co-create properly coded components that meet accessibility guidelines. This means we can launch products, apps and websites that begin at a much higher level of accessibility.

However, AI is only half the equation - humans will always need to provide the genuine usability and accessibility insights, in terms of context and all the nuances that only humans understand as AI cannot grasp the lived experience of your users. AI won’t always get it right, not every caption will be accurate, not all images will be correctly described, and not all suggestions will be what people want. So, as we move forward with AI let’s ensure we as humans take responsibility to implement good judgement, good user experiences and of course good overall accessibility.

---

If you would like test any of your digital content with real target users with accessibility needs, please feel free to get in touch by clicking on the 'Get in touch' button.