A structured governance framework for managing AI risks in data-driven transformations

Traditional risk frameworks fail to address the challenges AI introduces into change programs: probabilistic assessment breaks down under Knightian uncertainty, emergent system behaviour defies prediction, and human-in-the-loop oversight too often provides psychological rather than functional protection. Managing AI Risks in Data-Driven Transformations responds with the Human Firewall, an original governance framework that moves the term beyond its familiar cyber-security usage into a rigorous, three-pillar model of AI oversight, integrating ethical stewardship, institutional accountability, and behavioral assurance.

The book translates these pillars into auditable practice through a suite of practical instruments: the H-E-V-R risk assessment framework (Hazard, Exposure, Vulnerability, Response), the AI Ethics Risk Register, the Trust Impact Matrix, and the six-step Risk-Ethics Integrated Assessment (REIA) methodology, all aligned to the NIST AI Risk Management Framework, ISO/IEC 42001:2023, IEEE 7000-2021, and the EU AI Act. Governance models, templates, reflective questions, and sector-based case studies from financial services, healthcare, and the public sector demonstrate how to navigate technical and organisational risks in an integrated way.

The book also covers:

• Why conventional probabilistic risk assessment is structurally inadequate for AI, and how automation bias and selective adherence undermine human oversight
• A new AI risk taxonomy addressing algorithmic opacity, bias, model drift, and hallucination, with trust calibration as the governance objective
• Accountability architectures for AI Ethics Boards and stewardship committees, including the Three Lines of Defence adapted for AI and contestability rights under the GDPR and EU AI Act
• Capability building from classroom to boardroom: AI literacy, certification pathways, and an AI Risk Leadership Framework
• AI governance maturity model and adaptive governance strategies for regulation and technology that continue to evolve

Designed for project and program managers, risk and governance professionals, and board directors leading AI-driven change, this book also supports professionals pursuing ChPP, CEng, and related chartered and certification pathways. It functions as a graduate textbook for MBA, MSc AI, Data Science, and Technology Management programs, and as a resource for executive education and professional CPD.

Sayara Beg, CEng, ChPP, SFHEA, FIC, is a Chartered Engineer and Chartered Project Professional based at the Centre for Systems Engineering, University College London (UCL), specializing in AI governance, data ethics, and project delivery. A Senior Fellow of the Higher Education Academy and a Fellow of the Institute of Consulting, she holds patents in Data Science as a Framework and Citizen Data Science. Her career spans more than 26 years of leadership roles at HSBC, Deutsche Bank, Tesco, UBS, and Shelter UK, delivering AI-driven change initiatives focused on compliance, risk mitigation, and enterprise transformation. She is the originator of the Human Firewall governance framework presented in this book, and her peer-reviewed research on it was published in 2026.