Join us for a live update from the RAi UK research programme, where we’ll share progress and insights from our Keystone projects. This session will highlight key developments, findings, and emerging impact, as well as opportunities for collaboration across projects, followed by an opportunity for discussion and questions with the community.

PROBabLE Futures: Probabilistic AI Systems in Law Enforcement Futures

The research of PROBabLE Futures focuses on the use of Probabilistic AI in Law Enforcement and the delivery of an ‘operational-ready’ blueprint of an evidence-based, contextually informed and future-oriented framework for responsible AI in law enforcement. The PROBabLE Futures Responsible AI Checklist for Policing, developed in collaboration with the NPCC, has now been published by the College of Policing and circulated to all UK police forces. The checklist is featured alongside new technology guidance from the College, which outlines principles for using AI and emerging tech in ways that are transparent, ethical, and legally sound.

PHAWM: Participatory Harm Auditing Workbenches and Methodologies

This project examines fairness, bias, and accountability in predictive and generative AI across real-world contexts such as healthcare, education, media, and cultural heritage. Using diverse datasets, it explores how AI can reinforce inequalities and misinformation, and adopts a participatory approach to co-develop auditing methods with stakeholders. Key outputs include practical frameworks, “workbenches,” and training to support responsible AI evaluation across sectors.

AdSoLve: Addressing Socio-technical Limitations of LLMs for Medical and Social Computing

This project addresses the socio-technical limitations of Large Language Models (LLMs) to support their responsible use in sensitive domains such as healthcare and law. It develops new benchmarks, criteria, and evaluation methods to assess real-world performance, alongside novel technical solutions to improve reasoning, privacy, and bias mitigation. Through close collaboration with policymakers, industry, and domain experts, the project co-creates practical tools and frameworks for safer AI deployment. By focusing on applications like medical diagnostics, mental health, and legal support, it aims to bridge the gap between AI capabilities and real-world use.

RAi Cornerstone

An agile, responsive and creative programme of research, RAi’s Cornerstone complements and enhances the wider portfolio of funded research. Led by Prof Joel Fischer, and supported by Prof Kate Devlin and Dr Shoaib Ehsan as the RAi Cornerstone Leadership team, activities are co-created with stakeholders and refined through an iterative development process to ensure maximum impact and relevance.

Speakers:

• Professor Simone Stumpf, Professor of Responsible and Interactive Artificial Intelligence, University of Glasgow
• Professor Maria Liakata, Professor of Natural Language Processing, Queen Mary University of London
• Michele Sevegnani, Senior Lecturer, University of Glasgow
• Professor Joel Fischer, Cornerstone PI, Responsible Ai UK
• Professor Bashar Nuseibeh, Chair Research Pillar, Responsible Ai UK