The Equitable AI Research Roundtable (EARR): Towards Community-Based Decision Making in Responsible AI Development

Multidisciplinary expertise improves AI development: EARR demonstrates that bringing together experts from law, education, community engagement, social justice, and technology—beyond traditional computer science perspectives—expands the scope of what gets considered in AI development and helps identify "emergent ethical and social harms" that might otherwise be missed.

Community-based roundtables create accountability mechanisms: By establishing a coalition that includes nonprofits, NGOs, and universities alongside a large tech firm, EARR provides an institutional structure for critical external feedback on technology decisions, suggesting that responsible AI development benefits from built-in spaces for outside scrutiny and dialogue.

Shared learning reduces knowledge gaps between stakeholders: The paper emphasizes that EARR creates "a space for mutual learning," indicating that responsible AI requires not just experts talking at technologists, but genuine two-way engagement where different stakeholders develop deeper understanding of each other's concerns and constraints.

Defining equity requires ongoing dialogue, not predetermined solutions: Rather than imposing external definitions, EARR's workshops focused on collaborative work to "conceptualize equity and vulnerability as they relate to AI technology," suggesting that responsible AI development must involve continuous conversation about whose values and needs are being served.

Epistemic responsibility means acknowledging what you don't know: The paper highlights "epistemic curiosity and responsibility" as a key principle, implying that responsible AI development requires technologists and organizations to actively recognize the limits of their own expertise and remain open to perspectives from affected communities.