"Is Someone There? Do They Have a Gun": How Visual Information about Others Can Improve Personal Safety Management for Blind Individuals
Stacy M. Branham, Ali Abdolrahmani, William Easley, Morgan Scheuerman, Erick Ronquillo, Amy Hurst · 2017 · Proceedings of the 19th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '17) · doi:10.1145/3132525.3132534
Summary
This paper investigates an underexplored dimension of accessibility for people who are blind: the need for visual information about other people to manage personal physical safety. While decades of assistive technology research have focused on helping blind individuals navigate routes, landmarks, and building features, very little work has examined how blind people navigate around and in the presence of other people — particularly in threatening situations. The researchers conducted an online survey of 58 blind and low vision individuals and follow-up semi-structured interviews with 10 blind participants. The survey was initially framed around facial recognition technology for navigation, but participants consistently raised personal safety as a critical use case. All 10 interview participants shared stories of times when their safety was threatened by others, expressing a strong need for access to visual cues that could help them assess and respond to threats. The study builds on a small body of prior work, notably Ahmed et al.'s 2016 study on physical safety, security, and privacy for people with visual impairments. The researchers coin the term "Personal Safety Management" to describe informed, agential, and proactive participation in maintaining one's own physical safety, and argue this represents a significant gap in the ASSETS research community.
Key findings
Participants identified several categories of visual information they need for safety: whether someone is carrying a weapon, the facial expressions and body language of nearby people, whether a person in authority is wearing a proper uniform or badge, the number and proximity of people nearby, and clothing or appearance cues that might signal intent. The study revealed that existing assistive tools already impact safety perceptions — white canes and guide dogs can either increase safety (by deterring aggressors) or decrease it (by signaling vulnerability). Half of all interview participants wanted better tools for communicating with police, including locating officers, verifying their identity, describing criminals, and reading officers' facial expressions during tense interactions. Participants described safety challenges across multiple contexts: crowded public spaces, encounters with silent or undetectable people, escaping threatening situations, and managing distractions from unexpected contact. The paper identifies four broad research opportunity areas: answering visual questions to increase interpersonal safety, exploring the relationship between ATs and safety perceptions, facilitating communication with police, and designing AT that empowers proactive safety management.
Relevance
This paper opens an important and largely overlooked area of accessibility research. It challenges the field's predominant focus on wayfinding and obstacle avoidance by demonstrating that interpersonal safety is a primary concern for blind individuals navigating public spaces. The findings have direct implications for designers of computer vision and AI-powered assistive tools — these systems need to go beyond object and text recognition to provide information about people, their behavior, and potential threats. The study also highlights critical accessibility gaps in interactions with law enforcement, an area with life-or-death stakes for people with disabilities. For practitioners, the concept of Personal Safety Management provides a useful framework for evaluating whether assistive technologies address the full spectrum of navigation needs. The paper's caution about bias in facial recognition systems — which have historically misidentified people of color — remains highly relevant as these technologies mature.
Tags: blindness · personal safety · facial recognition · assistive technology · navigation · computer vision · interpersonal safety · law enforcement · violence