Responsible AI guidelines fall short at work? How responsible AI governance can benefit from the world of work perspectives

Artificial intelligence (AI) is becoming part of everyday work and people management, with potentially far-reaching consequences for employees, organizations, and labour markets. Yet the regulation of workplace AI is still evolving, leaving organizations and policymakers uncertain about what “Responsible AI” (RAI) means in practice. Governments, professional bodies, and organizations have produced a growing number of RAI guidelines. But do these guidelines help equip organizations and policymakers to govern AI in the workplace? This paper examines 183 organizational RAI guidelines to understand how they address work and employment issues. Our analysis identifies the main themes emphasized across the guidelines and reveals that work- and labour-related concerns rarely occupy a central position. It also suggests that, despite their proliferation, RAI guidelines remain centered on largely stable and abstract principles, with limited evolution in their normative focus. When workplace issues are mentioned, they are usually framed indirectly, most often through concerns about discrimination and employee safety. RAI frameworks therefore rarely confront the workplace directly, even though employment is one of AI's most immediate and consequential arenas. We argue that anchoring RAI governance in established governance mechanisms – including international labour standards and fundamental principles and rights at work – can help organizations and policymakers ensure that workplace AI addresses ethical concerns in the world of work and promotes decent work.