Transparency, Explainability and Accountability

The video centres on three principles that sit at the heart of ethical AI use in people practice: transparency, explainability, and accountability. Learners studying FREE08 and working towards AC 2.1 will find this video examines the well-documented "black box" problem in AI systems — where inputs and outputs are visible but the internal decision-making logic is opaque — and considers what that opacity means when AI is applied to recruitment shortlisting, promotion recommendations, and performance feedback.

Accountability is treated here as a distinct and genuinely complex question. The video addresses how responsibility for AI-driven outcomes can become diffuse across system developers, technology vendors, line managers, and senior executives, and why that diffusion poses a specific ethical risk in a people practice context. It considers the design of organisational frameworks that assign oversight responsibilities, establish authority to override AI recommendations, and ensure that someone is answerable when an AI-informed decision causes harm. These themes connect directly to the fairness, equality, and protected characteristics considerations named in the unit's indicative content.

A thread running throughout the video is the relationship between explainability and meaningful human oversight. The discussion covers what genuine oversight of AI recommendations looks like in contrast to passive ratification of automated outputs, and why the conditions under which managers review AI decisions — including time, information, and authority — determine whether that oversight is substantive. Real-world HR contexts, such as candidates being unable to receive reasons for rejection, are used to ground the ethical principles of fairness and dignity rather than treat them as abstractions.