Model Drift to Bias and Discrimination: The Many Risks of AI: Part 2

In Part 2 of this Lunchtime BABLing series on AI risk, Dr. Shea Brown, CEO of BABL AI, is joined again by Jeffery Recker to continue their lightning-round exploration of the real challenges organizations face when deploying AI. This episode dives deeper into critical concepts such as model drift, bias vs. discrimination, and growing explainability gaps in modern AI systems — especially as organizations increasingly rely on large language models and automated decision-making tools. Together, they discuss: -What model drift is and how organizations can detect and manage it -Why users (not just developers) should understand performance drift in AI systems -The important distinction between statistical bias and illegal discrimination -How bias can emerge even when demographic data isn’t explicitly used -The role of diversity of thought and structured risk assessments in uncovering AI risks -Why explainability is becoming harder as AI models grow more complex -The trade-offs between performance, trust, fairness, and regulatory compliance The conversation also explores broader questions around how AI is being used today, the limitations of “black-box” systems, and why validation, testing, and governance are becoming essential capabilities for organizations adopting AI at scale.

Lunchtime BABLing with Dr. Shea Brown · Shea Brown, Jeffery Recker