MEASURING AI AND GEN AI PERFORMANCE

When accuracy isn’t absolute – what’s next?
Developing new metrics for non-deterministic outputs

• Moving beyond traditional statistical measures
• Designing human-in-the-loop validation for GenAI outcomes
• Addressing instability and drift in AI model responses
• Exploring AI-to-AI validation and its circular challenges
• Sampling approaches to balance efficiency with reliability