AI Ethics & Policy Research

Algorithmic bias occurs when an AI system produces systematically different outcomes for different demographic groups. It arises from multiple sources: training data that reflects historical discrimination, feature selection that inadvertently encodes protected characteristics through proxy variables, and objective functions that optimise for aggregate accuracy at the expense of minority-group performance. Eliminating bias entirely is probably impossible because data is always historical and history is imperfect. The practical goal is to detect, measure, and mitigate bias to acceptable levels through diverse development teams, representative datasets, and ongoing monitoring.

The EU has adopted the most prescriptive approach, with binding rules, prohibited practices, and pre-market conformity assessments for high-risk systems. The US has favoured a sectoral strategy where existing regulators apply their authority to AI within their domains, supplemented by executive orders and voluntary frameworks like the NIST AI RMF. China has moved quickly on specific applications, regulating algorithmic recommendation systems, deep synthesis, and generative AI through targeted administrative rules. Each approach reflects different political traditions and priorities.

It depends on what is meant by transparency and for whom. Technical transparency, such as publishing model architectures and performance metrics, is straightforward but insufficient for non-expert audiences. Explainability for end-users means providing understandable reasons for specific decisions. For complex models, generating explanations that are simultaneously faithful to the model's reasoning and comprehensible to lay users remains an open research problem. Post-hoc explanation methods like SHAP and LIME are useful but approximate rather than reveal actual model behaviour. The field recognises that transparency is a spectrum, not a binary state.

Liability frameworks are still catching up with the technology. Under existing product liability law, manufacturers can be held liable for defects that cause physical harm. But the picture is murkier for non-physical harms such as discriminatory decisions or wrongful arrests. The EU AI Act introduces obligations on both deployers and providers, creating a chain of responsibility. The practical challenge is that AI systems often involve multiple actors, data providers, model developers, fine-tuners, and deployers, making it difficult to attribute causation cleanly.