This is an extended introductory tutorial on AI, with emphasis on Generative AI for a process safety audience. It starts by defining intelligence, artificial intelligence, and the major branches of AI, then surveys machine learning, supervised vs. unsupervised learning, NLP, vision, speech, robotics, and humanoid robots before going deeper into large language models. It explains what “ChatGPT” means, how GPT systems are generative and pre-trained, and how transformer-based models use attention, tokenization, and layered matrix operations to produce answers. The latter part of the deck focuses on practical concepts such as foundation vs. frontier models, instruction following, dialogue, tool use for deterministic calculations, zero/one/few-shot prompting, and in-context learning, with repeated examples tied to PSM such as P&ID extraction, MOC classification, flash-point retrieval, and relief-valve or LOPA-style calculations.
