4. Governing AI in Complex Organisations

This chapter examines how the rapid, organisation-wide adoption of artificial intelligence can unsettle an organisation’s purpose, values, and culture if introduced as a purely technical capability rather than a socio-technical change. It argues that organisations exist to serve society and depend on the quality of relationships with employees, customers, suppliers, and wider stakeholders; therefore, AI must be evaluated against the organisation’s mission and “lived” values, not just its espoused ones. The authors highlight recurring tensions—innovation versus accountability, speed versus trust, and “lightness” (convenience and acceleration) versus “weight” (meaning, responsibility, and governance)—and note that indiscriminate use of powerful generative models can undermine sustainability goals through resource consumption, while biased training data can threaten inclusion. They also warn that pervasive GenAI use may erode human agency and shift cultures from learning-oriented to transactional through the outsourcing of responsibility.To address these risks while realising value, the chapter proposes an AI-specific governance approach informed by established enterprise governance and the emerging AI management standard ISO 42001. It outlines four interlocking domains—strategic management, resource optimisation, risk management, and benefits realisation—emphasising short, iterative planning cycles and senior leadership involvement given AI’s fast-changing opportunity–risk landscape. The framework broadens governance beyond IT to include operational technologies and agentic AI, and stresses the need for clear roles and responsibilities, strong data and knowledge-base stewardship, and investment in foundational capabilities such as data integrity, learning agility, critical thinking, and cross-boundary collaboration. Inclusion is positioned as a practical risk mitigator against groupthink, and sustainability is treated as a non-negotiable constraint spanning environmental, social, and economic dimensions. Overall, the chapter frames effective AI adoption as a deliberate balancing act: enabling innovation while institutionalising the accountability required to keep organisations aligned with their purpose.