AI and the People Who Run the Plant: Safety, Maintenance and Readiness
18 November 2026 • 10-11:15 a.m.
Description: “Every talk in this session is about supporting a person rather than replacing one. Yang Ni of Purdue University Northwest presents research on AI-driven hazard recognition, addressing why models trained on general imagery struggle with mill heat, dust, smoke and glare, and how AI can move beyond detecting hazards to helping safety personnel decide what to do about them. Ed LaBruna, Ariel Gonzalez and Ignacio Lobato then present an AI agent for maintenance management validated across three steel production facilities, integrated with plant ERP systems, answering questions in natural language, and built with deliberate constraints to limit hallucinated output. Dan Simkins and Doogie Levine of ATiiD close by arguing that every AI initiative rests on four foundations, data, process, workforce literacy and willingness to change, failing at whichever is weakest. Attendees complete a short anonymous readiness self-assessment and leave knowing which foundation is their real constraint.”
Organized by: AIST Digitalization Applications Technology Committee