New standard to mitigate negative impacts of AI systems on people with disabilities

Author Sharlene Gandhi

Accessibility Standards Canada is asking the public for feedback about a new standard that seeks to ensure equitable access and development of artificial intelligence for people with disabilities. 

People with disabilities can often be categorized as ‘statistical outliers’ in the data that trains artificial intelligence systems. This standard calls for developers to mitigate these impacts and stop producing machine learning algorithms that further marginalize people with disabilities. 

According to the standard, AI must also work well with assistive technologies.

“Nothing without us” is a principle embedded in this standard—people with disabilities should be included in every part of the development of AI systems, from design to testing, said ASC. 

The draft version also says people with disabilities should have access to alternatives when faced with AI decision-making systems, such as human-made decisions. 

The standard is open for public review until May 5.

Tell us this made you smarter | Contact us | Report error

  • Sharlene Gandhi is the Future of Good editorial fellow on digital transformation.

    Sharlene has been reporting on responsible business, environmental sustainability and technology in the UK and Canada since 2018. She has worked with various organizations during this time, including the Stanford Social Innovation Review, the Pentland Centre for Sustainability in Business at Lancaster University, AIGA Eye on Design, Social Enterprise UK and Nature is a Human Right. Sharlene moved to Toronto in early 2023 to join the Future of Good team, where she has been reporting at the intersections of technology, data and social purpose work. Her reporting has spanned several subject areas, including AI policy, cybersecurity, ethical data collection, and technology partnerships between the private, public and third sectors.

    View all posts