The Perceptual Gap: Why we need Accessible XAI for Assistive Technologies

Posted on:April 13, 2026 at 12:00 AM

A simplified version of Fig. 1 from the paper. On the left, an example image that is classified as a flower is shown, but in reality is a flower on a computer screen loaded into an editing software. The user asks, is prediction correct? The abled user is able to tell that the visual explanation misses some context in the image and does not account for the software interface that's visible, while the BLV user cannot tell that.

Full Paper | Preprint

Abstract

Artificial intelligence systems are widely used by people with sensory disabilities, like loss of vision or hearing, to help perceive or navigate the world around them. This includes tasks like describing an image or object they cannot touch, reading documents, automatically captioning speech, and so on. Presently, models used for these tasks are based on deep neural networks and are thusly black boxes. Explainable AI (XAI) describes methods that can explain why a model gave the output it did. However, existing XAI methodologies are rarely accessible or designed with disabled users in mind. In this paper, we survey existing work in XAI with a focus on human-centered and accessibility-centered approaches or evaluations. We show that there is next-to-no XAI work that accounts for people with sensory disabilities, that many typical explanations are difficult for them to comprehend, and propose possible avenues for future work in Accessible Human-Centered XAI.

Credits

Shadab Hafiz Choudhury 1 2

1: Department of Computer Science and Electrical Engineering, University of Maryland, Baltimore County
2: Corresponding Author

Citation

@inproceedings{choudhury_perceptual_2026,
	title = {The {Perceptual} {Gap}: {Why} {We} {Need} {Accessible} {XAI} for {Assistive} {Technologies}},
	shorttitle = {The {Perceptual} {Gap}},
	series = {{CHI} {EA} '26},
	author = {Choudhury, Shadab H},
	month = apr,
	year = {2026},
	url = {https://dl.acm.org/doi/10.1145/3772363.3799105},
	doi = {10.1145/3772363.3799105},
	booktitle = {Proceedings of the {Extended} {Abstracts} of the 2026 {CHI} {Conference} on {Human} {Factors} in {Computing} {Systems}},
	publisher = {Association for Computing Machinery},
	address = {New York, NY, USA},
}