The "Who", "What", and "How" of Responsible AI Governance: A Systematic Review and Meta-Analysis of (Actor, Stage)-Specific Tools

Most responsible AI tools focus narrowly on technical roles: The review of over 220 tools found that the vast majority were designed for AI designers and developers during data and statistical modeling stages, leaving other critical stakeholders—including institutional leadership, deployers, end-users, and affected communities—without adequate support tools.

Significant gaps exist in the AI lifecycle: Tools are heavily concentrated in early technical stages (data preparation and model building) while neglecting crucial phases like defining the value proposition and deployment, meaning organizations lack guidance precisely when AI systems enter the real world.

Responsibility confusion undermines implementation: The paper identifies that responsibilities across the AI lifecycle are "ambiguously defined and assigned," leading to confusion and miscommunication among stakeholders—a fundamental governance problem that tools alone cannot solve without clearer role definition.

Institutional leadership and communities are underserved: Decision-makers at the organizational level and the communities potentially affected by AI systems have few tools to guide their participation in responsible AI governance, despite their importance to ethical implementation.

Current tools don't address the full governance challenge: Despite the existence of numerous responsible AI frameworks and tools, the research reveals that responsible AI implementation remains incomplete, suggesting that having more tools isn't enough without better coordination across stakeholder roles and lifecycle stages.