RelAItionship Building: Analyzing Recruitment Strategies for Participatory AI

Participatory AI requires intentional recruitment strategies, but current practices are poorly documented: Analysis of 37 AI projects revealed inconsistent documentation of how stakeholder groups are identified and recruited, making it difficult for the field to learn from successes and failures in engaging affected communities.

Recruitment outcomes depend heavily on researcher relationships and expectations, not just methodology: Interviews with five AI researchers showed that who gets involved in AI design is shaped as much by the relationships researchers build and their own goals as by formal recruitment processes, suggesting that "relationship-forward" approaches matter more than standardized procedures.

Structural barriers limit who can participate in AI development: The paper identifies that researchers face practical constraints—including time, funding, and institutional pressures—that affect which stakeholder groups they can realistically reach and engage, potentially excluding the most marginalized voices that participatory AI aims to include.

Current recruitment practices may not achieve equity goals: Despite researchers' stated intentions to center equity and empowerment, the study suggests that existing recruitment methodologies often fall short of genuinely including all relevant stakeholders, raising questions about whether participatory AI is reaching those most impacted by AI systems.

The field needs more reflexive and transparent recruitment practices: The authors recommend that AI projects adopt "relationship-forward" and reflexive recruitment methods, implying researchers should be more intentional about examining their own biases and documenting how they decide who to involve in AI design.