What is the intersection of AI and compliance? What about machine learning? Are you using ChatGPT? These questions are just three of the many we will explore in this cutting-edge podcast series, Compliance and AI, hosted by Tom Fox, the award-winning Voice of Compliance. Today Tom visits with Paul Welter, a German lawyer and former software engineer who co-founded Bayshore AI after research at Stanford Law School’s Codex on automating legal reasoning.
Welter explains Bayshore’s approach: extracting legal decision logic into deterministic, conventional code for predictable, explainable outcomes, while using large language models to gather scenario facts and handle vague legal terms. They discuss regulation as infrastructure for trust and economic activity and how today’s complexity creates bottlenecks that AI can alleviate by providing legal and compliance advice “in abundance.” Welter describes tailoring assessments to each customer’s policies and risk appetite, embedding them into workflows, and building trust through transparency, second-line control, and auditability for regulators. He highlights automation candidates (e.g., gifts/hospitality, third-party reviews) and advises teams to centralize request intake, measure frequency and effort, and automate iteratively.
Key highlights:
- Predictable Legal AI
- Regulation Enables Progress
- Encoding Policies Not Laws
- Building Trust in AI
- Explainability and Audits
- Agentic AI Use Cases
Resources:
- Paul Welter on LinkedIn
Bayshore