This intersective practical workshop will discuss the OCC’s framework for AI model governance and how to work within consumer protection laws. This session will offer practical steps to explain what your LLM is doing, how it is making decisions, and the assurance that the LLM is not making bias decisions.

Meanwhile, in an effort to show transparency, the State of California incoming law (2026) will requires developers of generative AI systems to publish detailed documentation about their training data.

• Defining what is, and what is not Artificial Intelligence (AI), and other definitions including Large Language Models (LLM), bias, hallucinations Translating model logic into plain English narratives that let exam teams trace inputs to outputs 
• Ensuring the explainability to the regulator and chain of thought for how your AI functions, and ensuring AI transparency for the regulators   Defending models by using decision logs and challenger tests so supervisors can verify fairness and control effectiveness 
• Applying the OCC model governance when you don’t know how the LLM is making decision Explaining how your organization’s AI tool or services works, what data was used to train it and how it was created  Proving fairness with documented metrics, thresholds, and monitoring triggers aligned to lending and underwriting rules so legal exposure is reduced 
• Protecting confidentiality in evidence packs by redacting customer data  
• Demonstrating safeguards against hallucinations and other invalid outputs