Cross-Checking

A verification strategy used by blind and low vision people to assess the reliability of AI-generated image descriptions by comparing outputs from multiple AI tools, taking photos from different angles, using non-visual senses, or consulting sighted individuals. BLV users have developed sophisticated cross-checking practices as workarounds for the unreliability of individual AI tools—for example, sending the same image to Be My AI, ChatGPT, and Seeing AI to compare descriptions and identify inconsistencies. While effective, manual cross-checking is time-consuming and impractical for everyday use, motivating the development of automated variation surfacing systems that systematically compare multiple model outputs.