7 months ago

ActInf GuestStream #096.1
Ron Itelman: "Cognitive Experience Design (CXD): Designing AI Systems for Trust"

A fundamental challenge in active inference is observing and verifying beliefs when designing intelligent systems composed of human and machine learners.This talk presents an approach that makes AI decision-making processes transparent and verifiable, similar to how git makes code changes visible.

As protocols for integrating LLMs with applications emerge (e.g., Anthropic's Model-Context-Protocol), we now have opportunities to combine deterministic and probabilistic components in a control systems approach. By introducing a clear design framework for capturing and verifying both human and AI intent, this talk introduces how to enable continuous, real-time belief updating with clear lineage.

Through practical examples and reference implementations, this talk shows how to create feedback mechanisms that capture problem-solving processes and enable real-time recommendations for knowledge and decision-making.

Active Inference Institute information:
Website: https://activeinference.institute/
X: https://x.com/InferenceActive
Discord: https://discord.activeinference.institute/
Donate: https://donate.activeinference.institute/
YouTube: https://www.youtube.com/c/ActiveInference/
Active Inference Livestreams: https://video.activeinference.institute/