Channelling, Coordinating, Collaborating: A Three-Layer Framework for Disability-Centered Human-Agent Collaboration

Researchers Lan Xiao and Catherine Holloway have published a new framework that fundamentally rethinks how AI agents should support people with disabilities. Their paper, "Channelling, Coordinating, Collaborating: A Three-Layer Framework for Disability-Centered Human-Agent Collaboration," argues that current AI accessibility tools are too narrowly focused on individual use—helping one person overcome a specific functional barrier. Instead, they propose AI should be designed to support the collaborative, interdependent ways people with disabilities already work, often accomplishing complex tasks through ability-diverse teams.

The framework consists of three distinct layers: Channelling (establishing shared informational ground across different abilities), Coordinating (mediating workflows between collaborators with different abilities), and Co-Creating (where AI contributes as a bounded partner toward shared goals). This approach extends the "agents as remote collaborators" vision by specifically centering the collaborative needs of people with disabilities, moving beyond solitary assistance to enable more effective team-based problem solving.

Grounded in established theories including the Ability-Diverse Collaboration framework, grounding theory, and Carlile's 3T framework, this research represents a significant shift in how we conceptualize AI's role in accessibility. The work has been accepted for presentation at the CHI '26 Workshop on Human-Agent Collaboration, indicating its relevance to both human-computer interaction and artificial intelligence research communities.