RAi Keystone Project: PHAWM – Participatory Harm Auditing Workbenches and Methodologies

WATCH THE FULL RECORDING HERE

In this webinar, Professor Simone Stumpf from the University of Glasgow introduces PHAWM – Participatory Harm Auditing Workbenches and Methodologies, a four-year, £3.4M RAi funded project.

This groundbreaking project brings together 25 researchers from seven leading UK universities, alongside 23 partner organisations, to address a critical challenge in AI: the lack of systematic auditing to assess the potential harms of predictive and generative AI.

Led by the University of Glasgow, with support from institutions such as the Universities of Edinburgh, Sheffield, and King’s College London, the consortium aims to develop methods to maximise the benefits of predictive and generative AI while minimising risks like bias and AI “hallucinations.”

This project will pioneer participatory AI auditing where diverse stakeholders without an AI background undertake audits of predictive and generative AI, either individually or collectively.

The predictive AI use cases in the research will focus on health and media content, analysing data sets for predicting hospital readmissions and assessing child attachment for potential bias, and examining fairness in search engines and hate speech detection on social media.

Professor Stumpf covered how to:

• Build novel participatory auditing workbenches for designing and deploying predictive and generative AI, targeted at diverse stakeholders
• Develop novel participatory AI auditing methodologies
• Embed participatory auditing in future AI deployment practices
• Extend the network of researchers in Responsible AI

The webinar was chaired by Professor Elvira Perez Vallejos, Chair Equities Pillar – RAi UK.

For more information on this project, visit RAi Keystone Projects.