Explaining decisions made with AI: A workbook (Use case 1: AI-assisted recruitment tool)

Explainability requires a multi-stakeholder approach: The guidance emphasizes that explaining AI decisions involves four key principles and must be designed with input from multiple perspectives—not just technical teams—suggesting that organizations need cross-functional teams including legal, ethical, and user-centered experts to implement effective AI explanation practices.

Different audiences need different types of explanations: The workbook identifies a "typology of AI explanations," indicating that a single explanation won't work for all stakeholders; recruiters, job applicants, regulators, and executives each require tailored explanations appropriate to their needs and level of technical understanding.

Explanation must be built in from the start, not added later: The framework emphasizes "explanation-aware design, development, and use," meaning organizations should plan for explainability during the initial design phase of AI systems rather than attempting to retrofit explanations after deployment—a critical consideration for responsible AI implementation.

Practical guidance exists and has been tested: The Alan Turing Institute and ICO's comprehensive guidance (published May 2020) represents "the most comprehensive practical guidance on AI explanation produced anywhere to date," and this workbook demonstrates that the principles can be applied to real-world scenarios like recruitment, making responsible AI explanation actionable rather than purely theoretical.

Recruitment is a high-stakes domain requiring particular attention to explanation: The choice to focus the workbook's case study on AI-assisted recruitment highlights that employment decisions significantly impact individuals' lives, making transparent and explainable AI systems especially important in this sector.