A report titled “To Reduce Disability Bias in Technology, Start With Disability Data,” published by the Center for Democracy & Technology (CDT), explores the pervasive issue of disability bias in technology and offers comprehensive recommendations for addressing this problem through the lens of disability data justice. The report begins by emphasizing the high stakes involved when people with disabilities interact with technology, particularly in areas such as employment, benefits, and healthcare. The authors argue that the increasing reliance on algorithmic systems in these areas has led to discriminatory outcomes, as these systems often fail to account for the unique experiences and needs of disabled individuals.
One of the core issues identified in the report is the inadequacy of current disability data sets. The authors discuss how these data sets are often non-representative, either because they do not include sufficient information about disabled individuals or because they are based on inaccurate definitions of disability. The report explains that the way disability is defined can vary significantly depending on the context—whether legal, medical, social, or personal identity—and these variations can lead to inconsistencies and inaccuracies in data collection. For instance, legal definitions like those found in the Americans with Disabilities Act (ADA) focus on impairments that limit major life activities, while the medical model views disability as a condition to be cured. In contrast, the social model of disability considers the role of societal barriers in disabling individuals, and the identity model embraces disability as an integral part of one’s identity. The report stresses that these differing definitions complicate efforts to gather accurate data, which in turn affects the effectiveness of technologies that rely on this data.
The report also highlights the impact of disability stigma on data collection. In many communities, disability is heavily stigmatized, leading individuals to avoid identifying as disabled. This reluctance is often exacerbated by internalized ableism, where individuals with disabilities may feel shame or denial about their condition. The stigma surrounding disability can thus result in significant underreporting in data collection efforts, further skewing the representation of disabled individuals in technology systems.
The authors of the report examine the exclusionary practices that are often inherent in data collection methods. They point out that traditional data collection approaches were not designed with disabled people in mind, which limits the generalizability of findings and can lead to biased outcomes. For example, data collection efforts frequently exclude environments such as group homes or correctional facilities, where a disproportionately high number of disabled individuals reside. Moreover, the tools and materials used in data collection are often inaccessible to those with disabilities, such as survey questionnaires that are not compatible with screen readers. These exclusionary practices are particularly concerning in the tech industry, where there is often a tension between the need to collect disability-related data and the pressure to minimize the collection of sensitive information.
The report goes on to discuss the broader implications of these data collection issues within the disability data ecosystem. The authors argue that the absence of comprehensive and representative disability data hampers efforts to identify and address bias in algorithmic systems. Without robust data on the experiences of disabled individuals, it becomes difficult to measure the impact of these systems on disabled communities or to develop effective interventions to mitigate bias. This lack of data not only perpetuates discrimination against disabled people but also exacerbates the marginalization of those who belong to multiple disadvantaged groups, such as disabled people of color or LGBTQ+ individuals with disabilities.
In response to these challenges, the report advocates for a disability data justice approach. This approach is rooted in the principle that people with disabilities must be involved in every stage of decision-making that affects their lives, particularly in the development and implementation of technology and AI policies. The authors offer several recommendations to achieve disability data justice. They emphasize the need to collect disability data in all demographic contexts where other types of data are collected, such as age, race, and gender. This would ensure that disability is recognized as a core demographic characteristic, rather than an afterthought.
The report also calls for data to be collected and stored in ways that respect personal and data privacy, ensuring that individuals’ information is used ethically and with their consent. Additionally, it underscores the importance of developing more inclusive methods of defining disability and collecting disability data, incorporating the perspectives of disabled individuals and experts with lived experience. The authors argue that practitioners should adopt a growth mindset regarding disability data, continuously improving their methodologies while recognizing the limitations of the data they currently have.
Furthermore, the report stresses the importance of including disabled people in the creation, deployment, and auditing of all technologies, particularly those that use algorithmic systems or biometric processes. The authors advocate for centering disabled leaders in the formulation of technology and AI policies to ensure that these policies adequately address the needs and concerns of disabled communities. Finally, the report recommends that data collection and storage practices be made accessible to individuals with disabilities, such as by ensuring compatibility with assistive technologies.
In conclusion, the CDT report provides a detailed examination of the problems with current disability data practices and offers a framework for addressing these issues through a disability data justice approach. By centering the experiences and expertise of disabled individuals, the report aims to promote the development of technologies that are more inclusive, equitable, and just.
