Accessibility & inclusion: AI for people with disabilities

AI continues to take over the technology space, but as we innovate, we need to be mindful of building AI that is usable by all people, including people with disabilities. We need to ensure the fabric of AI is based on the foundation of all human experiences, including temporary, permanent and situational exclusions. AI has the potential to play a huge role in empowering everyone, including people with disabilities. According to WHO, there are 1.3 billion people with disabilities globally, which equates to 1 in 6 people who identify with a disability. We all will experience some form of disability at some point in our lives – whether it’s a temporary injury like a fractured arm, or experiencing a situational exclusion with being deaf in loud areas.

There are a plethora of content creation generative AI tools like ChatGPT that are transforming how we create using technology. Developers need to account for multiple input modalities to ensure these are usable by a variety of users. One application of this principle is when ChatGPT introduced a customisation feature in July that considers custom instructions for all conversations moving forward. Thus, users will not have to repeat preferences or information in each conversation, enabling models to effectively reflect the diverse contexts and unique needs of each person. For instance, a developer who prefers to see examples in Python would add the preference one time, and the model remembers that preference going forward. This improves the user experience and makes this an accessibility feature for people with cognitive disabilities. In addition to typing the prompts, these AI tools should provide alternative input ways for users such as dictating. ChatGPT recently added support with a microphone button on their website, and other third party apps like BingAI allow the use of ChatGPT with voice.

AI will be invaluable in empowering users with disabilities to move towards a more contextual and interactive user experience. AI models should be trained on all human experiences, including people with disabilities, to understand the broader context of user and intent.

It has been shown that 56% of products designed with disability in mind are beneficial to all users. For example, developers are using GPT4-integrated technology to write code. This benefits developers with motor disabilities by reducing repetitive steps like adding multiple parentheses while writing code. With GitHub Copilot, developers with cognitive disabilities can write code more effectively; the Copilot provides scaffolding for new code and languages so as to support users by not forcing them to remember and recall all the different functions. For example, if a developer is attempting to write a function in a computer language they are unfamiliar with, the Copilot will support their experience by providing incontext suggestions that support their tasks through those learning structures. It is also fascinating that the system will adapt to the developer’s personal coding style over time and is a great way to learn new languages.

Another example is Beme.AI, which combines wearable technology with AI for people with autism to thrive through well-being, development tracking, and analytics. This application can be extended to varied healthcare and fitness use cases such as Fitbit and Apple Watch that use sleep-tracking data for health-related insights. IWill, one of the first digital mental health startups in India, aims to bring an AI-based cognitive behavioural therapy (CBT) model in English and Hindi through the IWill Care project. Mental health support will be valuable for all. Gen AI tools have the potential to break barriers for people with disabilities across industries, including the creative spaces of art and painting. Consider an artist who developed rheumatoid arthritis, or a painter who lost his arm in an accident. How might they continue to foster their passion for art? Artistic Gen AI tools like DALL·E, Canva AI and Jasper Art provide the ability for all artists to participate in the creative world through input prompts and images to generate high-quality art.


Nandita Gupta is an accessibility product owner & manager at Microsoft, Redmond, US


Get an chance to win ₹5000 Amazon Voucher by taking part in India's Biggest Habit Index! Take the survey here