Finding the Signal in the Noise: An Exploratory Study on Assessing the Effectiveness of AI and Accessibility Forums for Blind Users' Support Needs

A new study from researchers at Stony Brook University and Georgia Tech reveals significant shortcomings in how blind users access technical support through both traditional forums and modern AI tools. The paper, submitted to CHI 2026 and based on IRB-approved interviews with 14 blind participants, found that while accessibility forums and generative AI have become vital resources, both introduce substantial barriers that increase cognitive load.

Traditional forums overwhelm users with multiple overlapping topics, redundant content, and fragmented responses that require mental assembly. Meanwhile, generative AI tools like ChatGPT introduce new problems by producing unreliable answers including fabricated information, overly verbose guidance, and contradictory suggestions that fail to follow prompts. The study specifically noted that AI-generated responses often include "hallucinations" (fabricated information) and fail to provide concise, actionable guidance, forcing users to spend additional time verifying information.

The research provides concrete data showing how current support systems fail blind users who rely on these resources for learning about screen reader features, software updates, and troubleshooting computer-interaction issues. The team outlined specific design opportunities to improve reliability and reduce cognitive demands, emphasizing the need for more trustworthy support systems. This comes at a critical time as AI tools become increasingly integrated into accessibility workflows, yet their limitations for specific user groups remain poorly understood.