23.05.2026
Launch of SAINTS AI Safety Seminar Series
This week, we launched our monthly SAINTS AI Safety Seminar Series.
The inaugural seminar, led by our Director Ibrahim Habli, explored the interdisciplinary and collaborative perspectives that characterise SAINTS training, alongside the various conceptions and misconceptions surrounding AI safety research.
The question of which types of harms should be included or excluded within AI safety generated the most robust debate. The discussion was driven by these examples (from the AI Safety: Navigating the Expanding Landscape of Potential Harms publication):
A. A generative AI system that spreads fake news, eroding confidence in independent regulators
B. An AI-based sentencing system that produces racially-biassed recommendations
C. An autonomous car that misclassified stop signs and leads to collision
D. An AI-based clinical-decision support system that weakens trust in human clinical judgement
E. An AI-based health screening system that is at least as performant as current non-AIservices but underperforms for patients with dark skin
Following the seminar, SAINTS PhD researchers and academics had to engage in a debate of a completely different kind: which afternoon treat to tackle first, scones with jam and clotted cream or shortbread biscuits!
We hope to extend invitations to future seminars to our SAINTS partners and the wider community.
