← All reviews

Bridging the Gap: Towards Advancing Privacy and Accessibility

Rahaf Alharbi, Robin N. Brewer, Gesu India, Lotus Zhang, Leah Findlater, Yixin Zou, Abigale Stangl · 2023 · Proceedings of the 25th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '23) · doi:10.1145/3597638.3615653

Summary

This workshop proposal addresses a critical but understudied gap between the accessibility and privacy research communities. The authors argue that while all technology has privacy implications, accessibility tools carry particularly overlooked privacy risks — and conversely, mainstream privacy-preserving techniques are often inaccessible to disabled users. The paper outlines a one-day virtual workshop held at ASSETS 2023 designed to bring together accessibility researchers, privacy researchers, policy experts, disability community members, and activists to explore how to build technologies that are simultaneously accessible and privacy-preserving. The workshop is structured around speculative design activities inspired by Casey Fiesler's Black Mirror Writers Room exercise, where participants imagine future technologies that could create both privacy and accessibility harms. The background section highlights several concerning dynamics: disabled people in the United States are heavily surveilled by systems that determine access to healthcare, employment, and housing, with no avenues to contest or refuse such interventions. AI-powered accessibility tools like computer vision aids for blind people often lack transparency about data practices and impose significant privacy risks. The concept of "Forced Intimacy," coined by disability and transformative justice activist Mia Mingus, describes how disabled people are expected to share very personal information with non-disabled people just to access basic services — a dynamic that technology frequently reinforces rather than challenges.

Key findings

The paper identifies four key research questions driving the workshop: what privacy issues arise from emerging accessibility applications like generative AI and VR; how privacy-by-design and accessibility-by-design can be integrated into full-stack development; what unintended consequences AI-driven privacy preservation may have for disabled users; and how intersectional factors like race, gender, and sexuality interact with disability in privacy negotiations. The authors document specific accessibility barriers in existing privacy mechanisms: visual obfuscation techniques assume sighted users who can confirm their privacy is protected, making them inaccessible to blind and low-vision users; lock icons and visual credibility indicators are difficult for screen reader users; online privacy measures use complex layouts and jargon that exclude people with intellectual disabilities; and content-hiding methods are fundamentally misaligned with how non-visual users interact with technology. The workshop produced plans for a sustainable Discord community, an accessible Zine, and a call-to-action piece for ACM Interactions Magazine.

Relevance

This paper is important for accessibility practitioners because it surfaces a tension that is rarely discussed: the tools and systems built to improve accessibility can simultaneously erode privacy, and the mechanisms designed to protect privacy can create new accessibility barriers. For organizations deploying AI-powered accessibility features — such as image recognition, voice assistants, or automated captioning — this work highlights the need to evaluate privacy implications alongside accessibility benefits. The intersectional framing is particularly valuable, reminding practitioners that disabled people are not a monolithic group and that privacy concerns are shaped by overlapping identities including race, gender, and socioeconomic status. The concept of Forced Intimacy should inform how we design authentication, onboarding, and data collection flows to minimize unnecessary disclosure of disability status or personal information.

Tags: privacy · accessibility · intersectionality · AI ethics · blind and low vision · workshop · inclusive design · surveillance