Panel at Stockholm School of Economics: Involving AI in High-Stakes Decisions
Honored to join a panel at Stockholm School of Economics on Involving AI in high-stakes decisions in real organisations, hosted by the House of Innovation. The breakfast seminar brought together research and practice perspectives on how AI is reshaping organizational decision-making — from hiring to innovation funding — and where humans must remain in the loop.
The panel
- Lele Cao — Senior Principal AI Researcher, King AI Labs (Microsoft Gaming)
- Luca Henkel — Assistant Professor of Finance, Erasmus University Rotterdam (joining online)
- Moderator: Ruiqing “Sam” Cao — Assistant Professor, House of Innovation, SSE
The system, not just the model
A recurring theme in the discussion: the real challenge in deploying AI for high-stakes decisions isn’t model availability — it’s system design. Who is the tool actually for? What does it measure, and what does it implicitly de-prioritize? And where do accountability gaps open up when an agent acts on behalf of an institution?
On accountability
One point I came back to repeatedly:
If AI makes the decision, it becomes too easy for humans to disclaim responsibility.
Job applicants in recent research often prefer AI interviewers for their consistency and structure — but that same standardization becomes a liability when judgment, context, and seniority matter most. The conclusion isn't to remove AI from the loop, but to be deliberate about who the tool is for, what gets measured, and where humans must remain engaged.
Thanks to Sam Cao and the House of Innovation team for the invitation, and to Luca for sharing his research from Rotterdam. Read the SSE write-up of the event for more.