A curated collection of research, papers, and published resources organized across 11 categories, 20 RBPS elements, and every major AI discipline.
As interest in artificial intelligence (AI) accelerates across safety-critical industries, organizations face a growing challenge: distinguishing meaningful, process safety–relevant applications from general-purpose automation and hype.
psm.ai addresses this gap by organizing research and insights across the elements of CCPS Risk-Based Process Safety (RBPS), providing a structured, vendor-neutral view of how AI is being applied—while progressively expanding coverage across all 20 RBPS elements as relevant research and practical applications emerge.
Explore how AI is being applied in process safety disciplines →
The term “Retrieval Augmented Generation” or “RAG” comes up frequently in discussions when companies are trying to deploy GenAI.
The first paper, by a team from Air Products, discusses their success with using RAG to mine incident data.
For readers that want to understand how (and why!) RAG works, an approachable reference is the chapter on RAG in the textbook by Chip Huyen. The publisher, O’Reilly, stated that this was the most downloaded book of 2025, so many people are finding value in this text.
The final reference this week is the original paper by Patrick Lewis and the team at Facebook. This is technically very dense. Interestingly the paper was published two years before ChatGPT was released, so the RAG concept is independent of current frontier models.