AI has become increasingly popular the last few years, and the existing frameworks that financial services organisations have for managing AI models, model development and deployment aren’t compatible with the rapidly emerging generative AI models. If security is not considered up front, and learnings/discovery is closed to individual enterprises, it will take much longer to get up to speed.

The FINOS Simple AI Governance framework provides threat enumeration for LLM-based applications in financial services. By collaborating to identify common use cases and common pitfalls, the framework enhances security for everyone whose data is handled by these systems. The Working Group is in its early stages, and have some draft frameworks which are still being developed, but there are already some key lessons that have been identified.

This talk discusses the current state, and what common lessons there are to learn about it.