The Escalator Problem: Identifying Implicit Motion Blindness in AI for Accessibility

TL;DR

This paper highlights a critical limitation in current multimodal large language models, called Implicit Motion Blindness, which hampers their ability to perceive continuous motion, affecting their reliability in assistive applications for the visually impaired.

Contribution

It formally defines the Implicit Motion Blindness problem, analyzes its impact on user trust, and calls for a paradigm shift towards physical perception and new benchmarks.

Findings

Current models fail to perceive escalator direction accurately.

Implicit Motion Blindness undermines trust in assistive AI.

A call for new benchmarks focusing on physical perception.

Abstract

Multimodal Large Language Models (MLLMs) hold immense promise as assistive technologies for the blind and visually impaired (BVI) community. However, we identify a critical failure mode that undermines their trustworthiness in real-world applications. We introduce the Escalator Problem -- the inability of state-of-the-art models to perceive an escalator's direction of travel -- as a canonical example of a deeper limitation we term Implicit Motion Blindness. This blindness stems from the dominant frame-sampling paradigm in video understanding, which, by treating videos as discrete sequences of static images, fundamentally struggles to perceive continuous, low-signal motion. As a position paper, our contribution is not a new model but rather to: (I) formally articulate this blind spot, (II) analyze its implications for user trust, and (III) issue a call to action. We advocate for a paradigm shift from purely semantic recognition towards robust physical perception and urge the development of new, human-centered benchmarks that prioritize safety, reliability, and the genuine needs of users in dynamic environments.