The Artificial Intelligence for Inclusion:Strengthening Workforce Participation for Persons with Disabilities side event, held at the United Nations Headquarters. Credit: UN Web TV
By Oritro Karim
UNITED NATIONS, Jun 11 2025 – On June 10, the United Nations (UN) held a conference titled Artificial Intelligence for Inclusion: Strengthening Workforce Participation for Persons with Disabilities. This conference, which was organized by the Permanent Mission of Canada to the UN, featured a discussion by a panel of experts from various sectors, looking to shed light on the ways AI tools can be used to create inclusive workforces that maximize fairness and accessibility.
Since the mainstream adoption of generative AI systems in the early 2020s, many industries have been restructured. For many workers around the world, the implementation of AI tools have streamlined work processes, making once tedious tasks easier than ever before. Efficiency has been revolutionized, with many human workers being pushed to higher-level positions and creating a host of new jobs across numerous industries.
Despite these benefits, AI systems produce risks of unintentional bias and discrimination, particularly during the hiring process, limiting inclusivity and merit-based employment in the workforce. Additionally, AI systems that have been designed for able-bodied users have effectively shut out members of the disabled community.
Throughout this conference, the panel of experts discussed the methods through which AI systems can be transformed to benefit disabled individuals who have been disproportionately affected by job displacement and discrimination. Due to AI tools being a relatively new development in the global workforce, many industries lack the necessary structures to keep them from compromising a fair and equitable work environment.
“AI is transforming the way that we live, not just how we do business. Because of its rapid arrival to users, because of its regulation, free space is posing huge questions around inclusion, ethics, privacy, and some of our most fundamental institutions,” said Patty Hajdu, Canada’s Minister of Jobs and Families.
According to Dr. Jutta Treviranus, the Director and Professor at Inclusive Design Research Centre, OCAD University, the majority of AI tools used in the workforce use algorithms that create biases for those that are considered different from the vast majority. Treviranus states that roughly 90 percent of U.S. organizations rely on AI tools to hire and determine disciplinary action for employees. These systems are often trained to detect individuals who are perceived as outliers and cast them aside, creating “organizational monocultures” which harm the disabled community.
“Bias toward optimal patterns means bias towards difference. As AI gets better and better, it gets better at discrimination. Many of you are using programs that help with efficiency and help produce systems that eliminate anyone that is not optima,” said Treviranus. “We have created an international community that hopes to address statistical inequality and cumulative harm. In U.S.risk and impact assessments, anything that happens to an outlier is deemed to be statistically insignificant. We are facing statistical discrimination with these protections as well.”
Additionally, AI systems that are designed to support disabled individuals often only account for physical disabilities while neglecting individuals with intellectual disabilities. Disabled women are also disproportionately affected by data bias. Without considering these groups, AI systems are effectively working against promoting a diverse array of perspectives in the workplace, which in turn, hurt decision-making processes and innovation.
“AI can be a powerful equalizer and tool, only if it is developed with intentionality,” said A.H. Monjurul Kabir, the Senior Global Adviser and Team Leader at Gender Equality and Disability Inclusion at UN Women. “It is critical that (AI) does not deepen existing stigma, discrimination, and inequalities, especially for women and girls who face compounded layers of discrimination.”
“The unfortunate thing is that even if proportional representation was possible, AI will still rule against outliers and small minorities. It’s extremely difficult to get cluster analysis in disability…We need to look at what is done with that data and how it’s analyzed. Privacy protections do not work if you are highly unique. Differential privacy removes the pieces of data that are helpful to create AI data that will serve you,” added Treviranus.
Furthermore, disabled individuals around the world lack adequate access to AI-powered assistive technologies. With AI tools being implemented in all major sectors of industry, it is imperative that disabled workers are supplied with tools that streamline their work processes and keep their physical and/or intellectual conditions in mind.
“To some extent, addictive tech is a broken business model. The weight of the costs is on disabled individuals and public service…People with disabilities are paying far more for access that works poorly and is often broken,” said Treviranus. “AI using these life changing technologies usually work the least from people who need them the most. The farther you are from the average, the less it works. If the products you have are in a different language or your environment is poor, it will not work well,” she added.
According to Jürgen Dusel, theFederal Government Commissioner for Matters relating to Persons with Disabilities for Germany, workers with intellectual disabilities are currently receiving tablets that help them navigate their daily responsibilities in hotel jobs. Additionally, Hajdu states that in many parts of the world, disabled individuals face limited access to breathing technologies due to a lack of electricity in their environments.
To create comprehensive systems that benefit a wide spectrum of individuals, AI technology must be accessible for the most underserved communities. With disabled individuals persisting in every corner of the world, there must be reforms in accessibility to ensure that all people are afforded a fair chance to survive and succeed in their fields.
“The unexplored knowledge terrain is that entire area that faces intersectional barriers…..If you work with individuals who experience greatest barriers you will create a much more adaptive system with less need for help. In the long term, you are saving money and you don’t need to engage so many people. …I think there is an imperative to do this work we need to ensure these people creating this intelligence actually act intelligent,” said Treviranus.
IPS UN Bureau Report
