Philosophical Foundations of Mechanistic Understanding in AI. This project aims to design a comprehensive theoretical framework to understand the behaviour of AI systems mechanistically. By integratin

Description

Philosophical Foundations of Mechanistic Understanding in AI. This project aims to design a comprehensive theoretical framework to understand the behaviour of AI systems mechanistically. By integrating insights from philosophy, cognitive science, and computer science, the project will bridge the gap between the low-level mechanisms and observable behaviours of AI systems. Expected outcomes include the development of novel interpretability methods, significant academic publications including a monograph, an international conference, and enhanced collaborations across global academic and industry sectors. The project is anticipated to advance our theoretical understanding of AI systems and apply these insights to critical safety issues, thereby contributing to the responsible evolution of AI technology.. Scheme: Discovery Early Career Researcher Award. Field: 5003 - Philosophy. Lead: Dr Raphaël Millière