Source: Fortune
Summary
Business leaders at Fortune Brainstorm Tech discussed the challenges of managing AI risks, including accountability and transparency. Executives from May Mobility, Thomson Reuters, and SentinelOne shared their insights on designing systems that can regulate themselves and ensuring AI accountability. Techniques such as “LLM as a judge” and separate AI verification systems were highlighted as important for maintaining accountability. As AI performs more tasks, smart structures for verification will become increasingly critical.
Our Reading
The numbers tell one story.
Executives are worried about AI mistakes and are focusing on transparency and introspectability. May Mobility’s Edwin Olson wants to create systems that can explain their mistakes. Thomson Reuters’ Caitlin Halferty emphasizes the importance of validating AI output. SentinelOne’s Gregor Stewart notes that humans can’t verify all AI work, so agents will have to emulate human processes.
The strategy enters a familiar phase: relying on AI to regulate itself.
Author: Evan Null
