Issue link: https://iconnect007.uberflip.com/i/1503998
20 PCB007 MAGAZINE I JULY 2023 making a lot of progress in medical applica- tions, using AI to do pattern matching, diag- nosing conditions we never would have seen otherwise. That sort of application could play well in manufacturing. Oh, yeah. I have been keeping up with what's happening in the medical sciences and that looks interesting. At the same time, they're still cautioning that it takes a human to corrob- orate and diagnose; there's still some human expertise that's required to validate. For example, in medical imaging, being able to pick up cancers that even expert radi- ologists would have missed is interesting. I'm curious, though, what AI is seeing inside of that data. at method is not being shared with a radiologist. Alfred Korzybski, the father of general seman- tics, said, "e map is not the territory." What is the AI doing that makes it better able to interpret the map? Who's making the map better? Right. For me that has a direct correlation to what could be hap- pening in manufacturing. If, for example, you could do scans for defects on traces on PCBs, being able to catch them before they get out, but then you would be able to track that trace back , to ask where the defect emerged. Is it a problem in the manufac- turing? Is that a problem in source materials? Is it a problem with people on the production line, or the machines that are doing the assembly? Do you see any work underway in this space right now? In terms of using large AI models for interpret- ing data? Yes, not directly, but I imagine it's coming. We're in a gestational period for that. Experiments are being done within the cus- tomer organizations that I've worked with. How do you think that AI development is going to change TQM? I keep coming back to leadership. Leadership doesn't yet understand the theory needed to actually change first, and this could lead to depending on AI for answers to system ques- tions. For example, what I see here in Can- ada is a system of management used every- where, irrespective of AI adoption rates, that we imported decades ago from the United States. at management system makes us dependent upon reactionary interpretations of phenomena in our organizations. We don't see or think in systems and interac- tions. Now, inside the TQM space, I imagine it's like any- thing. It could make you more reliant on the AI to direct you, and less reliant on your own intelligence. If you haven't shied your thinking, you will be pre- sented with a lot of data points, and you won't nec- essarily be able to make the correct interpretations or ask the right questions, such as, "How do I improve a pro- cess? How do I manage a system?" What would be totally cool would be some- thing in AI that could actually help facilitate these questions: "How do I manage my orga- nization as a system?" or "How will this deci- sion affect other parts of the organization?" You want it to be able to provide guidance and waypoints so that you can say, "I'm con- sidering implementing this particular policy in this part of the plant. What are the impli- cations?" Chris, this has been a really interesting and informative conversion. Thank you so much. I agree. ank you, Nolan. PCB007 How do I improve a process? How do I manage a system?