Visual explanations for AI decisions: Fostering trust with transparency and control

This talk took place at a seminar organised by the Trustworthy and Responsible AI Lab at Sorbonne University, Paris. I introduced examples of how explainable AI can be approached from an algorithm-centred and human-centred perspective. Afterwards, I discussed how explanations often incorporate visualisations and presented examples from my own research into visual analytics, visualisation-supported transparency, and visualisation-supported control.

Diagram of algorithm-centred and human-centred XAI