Related Experiment Video
Updated: May 13, 2026

Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models
Published on: December 23, 2025
DescribePro: Collaborative Audio Description with Human-AI Interaction
Maryam S Cheema1, Sina Elahimanesh2, Samuel Martin1
1Arizona State University Tempe, Arizona, USA.
DescribePro enhances audio description (AD) creation for blind and low vision (BLV) users by combining AI efficiency with human creativity. This collaborative system allows iterative refinement and community input, improving AD quality and accessibility.
Area of Science:
- Human-Computer Interaction
- Accessibility Research
- Artificial Intelligence in Media
Background:
- Audio description (AD) is crucial for making video content accessible to blind and low vision (BLV) users.
- Current AD creation faces a trade-off between human precision and AI efficiency.
- Existing tools lack robust support for collaborative refinement and diverse narrative exploration.
Purpose of the Study:
- To introduce DescribePro, a novel collaborative system for authoring audio descriptions.
- To enable iterative refinement of AI-generated ADs using multimodal large language models and manual editing.
- To facilitate community collaboration for exploring diverse narrative styles and improving AD authorship.
Main Methods:
- Development of DescribePro, a collaborative AD authoring system.
- Integration of multimodal large language model prompting for AI-assisted description refinement.
- Implementation of community features for forking, editing, and versioning ADs.
- Evaluation with 18 describers (9 professionals, 9 novices) using mixed methods.
Main Results:
- AI support in DescribePro reduced repetitive tasks for AD describers.
- The system helped professionals maintain stylistic choices and eased cognitive load for novices.
- Collaborative features demonstrated potential for customization, version control, and training.
Conclusions:
- DescribePro effectively balances AI efficiency with human creativity in AD authoring.
- Collaborative, AI-assisted tools can significantly enhance and scale the production of high-quality audio descriptions.
- The system shows promise for improving accessibility and supporting diverse narrative approaches in AD.
Related Concept Videos
Automatic Processing and Automatic Social Behavior
Auditory Perception
Auditory Pathway
When viewed cross-sectionally, the cochlea reveals the scala vestibuli and scala tympani flanking the...
Impression Management Techniques IV: Altercasting
Hearing
Non-equilibrium in the Cell