ImageAlly: A Human-AI Hybrid Approach to Support Blind People in Detecting...
summary
ImageAlly is a tablet/phone app that uses a human-AI hybrid pipeline: an automated computer-vision screener flags potential private content, then users optionally send the image to a trusted sighted ally for redaction. An exploratory study with 20 participants (10 blind people paired with 10 allies) finds the system empowers blind users to prevent privacy leaks, catches edits allies miss (example: leftover credit-card digits), and scores highly on usability (SUS 86.2 / 84.2).
AI-written summary · about
related talks
05mWebAlly: Making Visual Task-based CAPTCHAs Transferable for People with Visual
50mAI as an A11y Ally
01h 00mAI and alt text
34mAI and Accessibility: the Good, the Bad, and the Bollocks
23mCooperative Privacy and Security: Learning from People with Visual Impairments
45mBe My Eyes - Leveraging Generative AI to help the blind and visually impaired people
11mImago Obscura: An Image Privacy AI Co-pilot to Enable Identification and Mitigation of Risks
21mA Face Recognition Application for People with Visual Impairments: Understanding Use Beyond the Lab