ChatGPT has proven that the possibilities for generative artificial intelligence are endless—even menacing, if you buy into the dystopian “Matrix”-style scenarios cooked up by commentators for society as a whole—but the possibilities of “gen-AI” for manufacturers are benign and, in fact, enormously beneficial for industry, especially for generating instructions, automating processes, and helping to avoid and recover from costly dreaded downtime, two experts emphasized during a recent Smart Industry-hosted online event.
SI’s Fall Insight five-part series of hourlong webinars concluded on Oct. 6 with a discussion among Michael MacKenzie, general manager of Industrial IoT and Edge Services for Amazon Web Services, Ibrahim Al-Syed, director of digital manufacturing at Celanese, and SI Editor-In-Chief Robert Schoenberger about how manufacturing businesses can capitalize on trending by often misunderstood gen-AI. The discussion, which was sponsored by Advanced Technology Services, Seeq and GE Digital, is available for viewing.
At AWS, MacKenzie helps customers in enterprise security, manufacturing, oil and gas, food and beverage, smart grid, and intelligent buildings connect and manage their critical operations from the cloud. Before coming to AWS in 2019, MacKenzie was the global VP of the EcoStruxure IoT Platform for Schneider Electric, a global automation and energy management company specializing in industrial IoT.
Al-Syed has spent over 13 years in downstream petrochemical sector, chemicals manufacturing across continents ranging from U.S. to Canada to Singapore. His passion is business transformation leveraging digital and data in leading reliability and operational excellence initiatives. He leads the digital transformation program for global manufacturing at Celanese.
“This is a new era," moderator Schoenberger pronounced to kick off the Oct. 6 discussion with SI, "with ChatGPT reaching 100 million users much faster than any other platform in internet history.” ChatGPT is the AI chatbot that relies on a large language model, or LLM, trained to mimic human writing by processing a large database of digitized books and writings and analyze how words are sequenced together.
“Could we see a world where generative AI could take your design data and create those documents, those drawings. That was never possible before. We see those possibilities … as a game-changer,” Al-Syed said, noting AI's obvious applications in building computer-assisted drawings of production machinery and infrastructure.
Gen-AI and predictive maintenance: Game-changer for downtime avoidance
The obvious application for gen-AI in manufacturing is predictive maintenance, all Oct. 6 webinar speakers agreed.
MacKenzie said data from sensors all over industrial equipment can be used in combination with gen-AI to monitor how the equipment is performing and changing over time as it ages and performs over the years, monitoring for cycles when maintenance might be necessary to keep functioning. In combination with IoT technologies, gen-AI can "flag possible preventive maintenance procedures to follow."
At one point, Al-Syed presented a scenario where gen-AI would be useful from a maintenance perspective, a situation where a pump is suffering from low flow and gen-AI can be trained to sort through standard operating procedures (SOPs) and give a technicians some insights on when to check for that malfunction. More generally, gen-AI can answer these maintenance questions and more: “What is the highest temperature of this equipment? This pump is about to fail, can I have all the relevant data to troubleshoot? I just finished a compressor overhaul, can you summarize all the findings and create a maintenance report out of it?”
Are there uses for gen-AI outside of maintenance, to make the job easier or more productive for those people who are operating machinery or the supervisors or the plant managers, Schoenberger asked.
Yes, MacKenzie answered, citing one example where a gen-AI "assistant" could be built into plant equipment to aid operators. "How do I do this? How do I do that? I need to unblock a chute, what's the SOP for that? It creates a very quick, very productive interface for someone to use."
Further along in the webinar discussion, Al-Syed observed: "We now have a significant amount of opportunity to use gen-AI, automating processes, getting effective decisions faster. People sift through documents, look under table for things now. The potential for their answers to be one click away is exciting."
MacKenzie added: "I’m excited to see what customers will do with the technology. This is a game-changing time-saver. If you’re an operator, get an SOP faster. How is downtime quickly resolved? This could be the answer."