Position Paper Proposes Argumentation as Foundation for Evaluative AI

A new position paper titled 'Towards an Argumentative Foundation for Evaluative AI' has been published on arXiv, advocating for computational argumentation as the formal basis for Evaluative AI (EAI). EAI is a recently proposed approach to AI decision support that, instead of offering a single recommendation, presents competing hypotheses with supporting and opposing evidence. The authors argue that argumentation is particularly suitable for making EAI explainable and contestable, and they outline a long-term research agenda aimed at developing distributed and human-centred EAI systems. The paper is categorized under Computer Science > Artificial Intelligence and was submitted on arXiv with the identifier 2608.07473. The abstract emphasizes the need for a formal, computable foundation for EAI that facilitates human oversight and contestability. The paper also mentions arXivLabs, a framework for collaborative experimental projects, and highlights arXiv's commitment to values such as openness, community, excellence, and user data privacy. The research is positioned as a foundational step towards more transparent and accountable AI decision-making systems.