Paper: Better Together: Human Oversight and Fairness in the EU AI Act

As discussions around the governance of artificial intelligence (AI) continue to evolve, a new paper published in the Cambridge Forum on AI Law and Governance makes a timely and thought-provoking contribution to the debate: “Better Together? Human Oversight as Means to Achieve Fairness in the European AI Act Governance”

Read the full paper here

As the European Union finalizes the implementation of the AI Act, one of its most contested and consequential elements is the requirement for human oversight in high-risk AI systems. This paper offers a critical yet constructive analysis of how human oversight is being positioned within the Act, highlighting both its potential and its current limitations.

What makes this paper especially valuable is its nuanced framing: rather than treating human oversight as a procedural safeguard alone, the authors argue for its role as a substantive enabler of fairness. This shift in perspective reframes oversight as an active governance mechanism, capable of aligning AI system behavior with societal values and legal norms.

Drawing on legal scholarship, governance theory, and applied AI ethics, the paper:

• Reviews how human oversight is currently defined and operationalized in the EU AI Act;

• Identifies gaps in the current regulatory framing, particularly around fairness and accountability;

• Proposes a richer conception of oversight that goes beyond technical correction and integrates human judgment, contextual understanding, and participatory deliberation.

For policymakers, researchers, developers, and practitioners working on AI governance, this paper serves as both a warning and a guide: if we reduce oversight to a checkbox or last-minute intervention, we risk undermining both fairness and trust. But if we invest in designing oversight mechanisms that are context-aware, inclusive, and value-sensitive, we move closer to AI systems that serve society, not the other way around.

In an era where AI decisions increasingly impact lives, this paper reminds us of a fundamental truth: fairness is not a feature to be coded, but a process to be governed, by humans, with care.

📖 Access the full paper: Better Together? Human Oversight as Means to Achieve Fairness in the European AI Act Governance