Be My AI Responses Too Verbose for Blind Users

A new Maryland study finds users must work around overly long, essay-like responses to get the practical information they need.

Updated on Oct. 23, 2026 in Artificial Intelligence

Be My AI Responses Too Verbose for Blind Users

Live Poll

Do you trust current AI tools to provide the practical information you need for daily tasks?

A University of Maryland study titled Say It My Way found that the Be My AI application frequently provides overly descriptive responses to blind users. The research analyzed 418 interactions and concluded that the AI lacks the conciseness required for practical, real-world tasks.

Why it matters

The findings highlight a critical usability gap where AI design prioritizes exhaustive image descriptions over the specific, goal-oriented needs of blind users. This misalignment forces individuals to employ manual workarounds to navigate daily activities efficiently.

Researchers observed that AI responses averaged 10 times the length of user questions, with conversations reaching up to 21 exchanges. Half of the 11 study participants cited the AI's initial full-image descriptions as a significant hindrance to their immediate goals.

The players

University of Maryland

A public research university in Maryland that conducted this study on assistive AI usability.

The details

The University of Maryland study assessed interaction patterns by observing 11 blind participants over a 10-day period. Participants often had to break complex tasks into smaller questions to bypass the AI's tendency toward verbose, essay-like responses rather than actionable guidance. The researchers identified technical barriers in image framing and camera positioning, noting that the AI also struggled with accurately estimating spatial distances and temporal changes.

The Say It My Way study was published on October 23, 2026.

Data Centers
Oct. 21
Cox Calls for Moratorium on Maryland Data Centers

The Tech Race

This study underscores the growing divide between general-purpose AI capabilities and the specialized UX requirements for assistive technology. It forces a recalibration of current development roadmaps that prioritize dense visual data over user-centric, concise interaction design.

Blind users currently utilizing Be My AI must manually decompose complex requests into smaller segments to avoid overly verbose feedback. These findings suggest that until developers implement context-aware response tuning, the app remains less efficient for high-priority navigation.

The takeaway

This study provides a clear benchmark for AI developers to optimize response length for accessibility applications. Future research will likely focus on whether generative models can be tuned to prioritize intent-based brevity over general-purpose description.