Artificial Intelligence (AI)‐based systems are widely employed nowadays to make decisions that have far‐reaching impact on individuals and society. Their decisions might affect everyone, everywhere, and anytime, entailing concerns about potential human rights issues. Therefore, it is necessary to move beyond traditional AI algorithms optimized for predictive performance and embed ethical and legal principles in their design, training, and deployment to ensure social good while still benefiting from the huge potential of the AI technology. The goal of this survey is to provide a broad multidisciplinary overview of the area of bias in AI systems, focusing on technical challenges and solutions as well as to suggest new research directions towards approaches well‐grounded in a legal frame. In this survey, we focus on data‐driven AI, as a large part of AI is powered nowadays by (big) data and powerful machine learning algorithms. If otherwise not specified, we use the general term bias to describe problems related to the gathering or processing of data that might result in prejudiced decisions on the bases of demographic features such as race, sex, and so forth.This article is categorized under: Commercial, Legal, and Ethical Issues > Fairness in Data Mining Commercial, Legal, and Ethical Issues > Ethical Considerations Commercial, Legal, and Ethical Issues > Legal Issues

In this work, we aim to classify nodes of unstructured peer-to-peer networks with communication uncertainty, such as users of decentralized social networks. Graph Neural Networks (GNNs) are known to improve the accuracy of simple classifiers in centralized settings by leveraging naturally occurring network links, but graph convolutional layers are challenging to implement in decentralized settings when node neighbors are not constantly available. We address this problem by employing decoupled GNNs, where base classifier predictions and errors are diffused through graphs after training. For these, we deploy pre-trained and gossip-trained base classifiers and implement peer-to-peer graph diffusion under communication uncertainty. In particular, we develop an asynchronous decentralized formulation of diffusion that converges to centralized predictions in distribution and linearly with respect to communication rates. We experiment on three real-world graphs with node features and labels and simulate peer-to-peer networks with uniformly random communication frequencies; given a portion of known labels, our decentralized graph diffusion achieves comparable accuracy to centralized GNNs with minimal communication overhead (less than 3% of what gossip training already adds).

Participatory design (PD) and co-creation (Co-C) approaches to building Artificial Intelligence (AI) systems have become increasingly popular exercises for ensuring greater social inclusion and fairness in technological transformation by accounting for the experiences of vulnerable or disadvantaged social groups; however, such design work is challenging in practice, partly because of the inaccessible domain of technical expertise inherent to AI design. This paper evaluates a methodological approach to make addressing AI bias more accessible by incorporating a training component on AI bias in a Co-C process with vulnerable and marginalized participant groups. This was applied by socio-technical researchers involved in creating an AI bias mitigation developer toolkit. This paper’s analysis emphasizes that critical reflection on how to use training in Co-C appropriately and how such training should be designed and implemented is necessary to ensure training allows for a genuinely more inclusive approach to AI systems design when those most at risk of being adversely affected by AI technologies are often not the intended end-users of said technologies. This is acutely relevant as Co-C exercises are increasingly used to demonstrate regulatory compliance and ethical practice by powerful institutions and actors developing AI systems, particularly in the ethical and regulatory environment coalescing around the European Union’s recent AI Act.

To help developers discover libraries suited to their software projects, automated approaches often start from already employed libraries and recommend more based on co-occurrence patterns in other projects. The most accurate project–library recommendation systems employ Graph Neural Networks (GNNs) that learn latent node representations for link prediction. However, GNNs need to be retrained when dependency graphs are updated, for example, to recommend libraries for new projects, and are thus unwieldy for scalable deployment. To avoid retraining, we propose that recommendations can instead be performed with graph filters; by analyzing dependency graph dynamics emulating human-driven library discovery, we identify low-pass filtering with memory as a promising direction and introduce a novel filter, called symmetric partially absorbing random walks, which infers rather than trains the parameters of filters with node-specific memory to guarantee low-pass filtering. Experiments on a dependency graph between Android projects and third-party libraries show that our approach makes recommendations with a quality and diversification loosely comparable to those state-of-the-art GNNs without computationally intensive retraining for new predictions.