Vertical Standardisation for High-Risk AI Systems under the EU AI Act: A Domain-Specific Framework for Algorithmic Hiring

As the EU AI Act mandates strict compliance for high-risk AI systems, including those used in recruitment, there is an urgent need for technical standards that translate legal requirements into actionable engineering practices. While horizontal standards provide general guidance on AI management and risk, they often fail to capture the unique, multi-stage complexities of hiring processes—such as job advertising, CV screening, and candidate ranking. This paper argues that a 'vertical' approach, tailored specifically to the recruitment domain, is essential to effectively mitigate risks like algorithmic bias and discrimination.

The authors introduce a framework that maps the high-level requirements of the EU AI Act (such as data governance, transparency, and human oversight) to concrete, actionable recommendations for recruitment systems. Unlike traditional standards that focus primarily on product safety or technical performance, this framework integrates fundamental rights protection directly into the development lifecycle. It emphasizes the need for fairness-aware data governance and post-deployment monitoring, ensuring that systems are not only technically sound but also socially responsible throughout their entire operational life.

The proposed framework is designed to be flexible, allowing developers to implement the recommendations using various methods and tools. By focusing on the specific challenges of ranking-based recruitment systems, the authors provide a roadmap for stakeholders—including developers, auditors, and HR professionals—to navigate the regulatory landscape. This work serves as a critical contribution to the ongoing standardisation efforts in Europe, aiming to bridge the gap between abstract legal mandates and the practical realities of AI-enabled hiring.