Why does it matter for society?

The aims of this workshop were to: 

• Identify public experiences of AI in decision-making, focussing on healthcare. 

• Explore public questions, concerns and aspirations about the reliability of AI in decision-making and the data used to build models and AI.  

• Provide societal context for the project, helping us understand what the public interest in research culture might look like

Session one: Data , Models and AI in Decision-Making

Marc Boubnovski-Martell, PhD student in the Department of Surgery and Cancer at Imperial, introduced the ways data, models and AI are used in decision-making in areas of healthcare such as diagnosis, treatment, drug discovery and planning. He discussed the role of researchers in developing AI tools. 

Session two: What Make an AI Trustworthy?

Zeynep Engin, Chair and Director at Data for Policy CIC, and Senior Researcher at University College London, opened up the black box of AI, discussing trustworthiness and what factors researchers, practitioners and developers should consider before a model or an AI is used to support a clinical decision, including openness, transparency, explainability and reliability. 

Session three: Challenges of Reliability in AI 

Raquel Iniesta, Senior Lecturer at King's College London, introduced the concept of reliability, why it matters and what can be done to maximise it. What role do governance frameworks play? Reliability is critical to understanding whether an AI tool is fit for purpose. What factors affect reliability? It might be accurate, but is it accurate enough? What role can different stakeholders play?