3rd December 2024
On December 3rd, we celebrate the International Day of Persons with Disabilities, an international observance promoted by the United Nations since 1992 to raise awareness about disability issues and mobilize support for the dignity and fundamental rights of individuals with disabilities.
In the workplace, however, people with disabilities continue to face significant barriers. EU-level data consistently highlights the significantly lower employment rates among persons with disabilities compared to their non-disabled peers. Recent figures reveal that the employment gap — the difference in employment rates between persons with and without disabilities — remains steady at around 24 percentage points, with even wider disparities for women and young persons with disabilities.
Research from the European Disability Forum (EDF) on Digital Skills, Accommodation and Technological Assistance for Employment shows how gaps in awareness, access to reasonable accommodation, digital skills’ development, and assistive technology provision all contribute to the persistent employment gap for persons with disabilities. For instance, many assistive devices are designed specifically for work-related activities, yet often remain unacquired due to employers’ general lack of awareness about assistive technologies. Additionally, many employers are unfamiliar with the range of assistive options available and how these technologies could benefit employees with disabilities.
EDF’s findings reveal that approximately 3 in 4 employers in the European Union and the United Kingdom are unsure whether their employees even use assistive technology. Additionally, 81% of surveyed employers lack policies for acquiring these technologies, and 77% are unaware of the legal framework that supports the acquisition of assistive technology. This lack of knowledge and policy creates significant barriers, leaving many workers with disabilities without the essential support they need.
Moreover, as workplaces increasingly embrace new technologies like AI systems, these disparities risk being exacerbated. Currently, AI systems are used in various aspects of the workplace, from automating recruitment processes and employee training to enhancing productivity through task management and decision-making tools. However, these technologies are likely to reinforce existing biases and discrimination if not designed, developed, and implemented with diversity and inclusivity at the forefront. Additionally, their use raises significant concerns about transparency, accountability, and data protection, risks that are especially heightened for individuals from historically oppressed groups. Without careful attention to fairness and accessibility, AI systems risk deepening workplace inequalities instead of creating a more inclusive environment.
This challenge is particularly pronounced for people with disabilities, who often face disproportionate challenges in accessing and mastering digital skills, placing them at a further disadvantage in the ongoing push for workplace digitalisation and automation. Despite the well-established link between digital skills and employability, research on the digital skills of persons with disabilities remains scarce at the EU level. For example, while data indicates that 53.92% of people in the EU had basic or above-basic digital skills in 2021, there is no disaggregated data on the proportion of persons with disabilities possessing similar skills, making direct comparisons impossible.
However, the problem extends beyond skill gaps and is frequently rooted in a lack of awareness and engagement by employers and AI developers. AI systems are more prone to errors when they encounter individuals whose physical features or communication styles fall outside the “norms” used to train these models. This reflects a broader problem highlighted by the Organisation for Economic Cooperation and Development (OECD): AI datasets are often unrepresentative of the disabled community, resulting in ableist biases that exacerbate workplace inequalities. These biases are further compounded by the frequent absence of essential accessibility features in AI systems, such as compatibility with screen readers or alternatives for individuals who cannot use speech, effectively excluding people with disabilities from full participation in the workforce.
The OECD also points out that many AI innovations aimed at assisting individuals with disabilities struggle to move beyond the prototype stage due to limited funding for research and development, challenges in recruiting qualified AI talent, and insufficient training for developers in accessibility. Another significant barrier to adopting accessible AI solutions is the lack of inclusion of persons with disabilities and users in the development process. This typically leads to AI solutions that do not address real needs and become impractical when not integrated into existing policies, stakeholder frameworks, and support systems. Furthermore, the implementation of these solutions is hindered by developers and decision-makers overlooking the generally lower IT skills of intended users, as well as the lack of interoperability between new AI solutions and existing hardware and assistive devices. Finally, the vast amounts of personal data processed by AI systems increase the risk of re-identification or unwanted disclosure of disability status, particularly for individuals with invisible disabilities.
Despite all of these challenges, AI systems are not always inherently problematic. In fact, AI-powered assistive technologies — such as voice-to-text software — are designed to break down barriers and enable people with disabilities to perform tasks more easily and independently. These technologies should be encouraged and incentivized in workplaces by both the employer and the State.
For AI systems intended for general use, diversity, and inclusivity must be a core priority from the outset. This requires involving people with disabilities and the organisations that represent throughout the development, validation, and implementation of the technologies. With their lived experience, these individuals provide invaluable insights into the challenges they face and the solutions they need. AI developers and employers must engage with them to understand their expectations, needs, and preferences, thus bridging the knowledge gap. By mainstreaming accessibility features and ensuring that AI tools are developed with everyone in mind, we can create more equitable and inclusive workplaces. However, many AI systems are still designed based on the typical functionality of individuals without disabilities, often reducing their effectiveness for users with disabilities, particularly those with intellectual disabilities.
Looking ahead, AI brings hope for improved digital support for persons with disabilities in the workplace. The availability of certain AI-powered assistive technologies — some of which are free or relatively affordable — adds promise by making these devices more accessible to both employers and individuals. However, many AI systems remain only partially accessible and lack the adaptability to learn and develop in ways that meet the diverse needs of users with disabilities.
As we observe this year’s International Day of Persons with Disabilities, let’s use this moment to push for AI that works for everyone. Technology should be a tool for inclusion, not exclusion. And we have the power to make that a reality — but only if we centre accessibility and diversity at the forefront of AI systems design and adoption.
