Podcast: Generative AI may be trendy, but manufacturers aren’t using it
Dennis Scimeca is the senior editor for technology at IndustryWeek. In his current position, he covers a range of topics, including how innovations in the manufacturing sphere are helping companies improve their competitiveness and their profits. He shines a light on the latest and greatest industrial technology, including vision systems, machine learning/artificial intelligence, virtual and augmented reality, and interactive entertainment. Dennis recently spoke with IndustryWeek editor in chief Robert Schoenberger about the results of the brand’s 2024 technology survey and how this information will impact the industry.
Below is an excerpt from the podcast:
IW: Every year, we ask our readers and manufacturing leaders what technologies they are using, what they think of various systems, what the state of technology is in the manufacturing world, and where they're expecting growth. And Dennis, to put it really bluntly and succinctly, when he was looking at this year's answers, said IIoT means nothing and no one uses generative AI. So Dennis, no one uses generative AI?
DS: Not for manufacturing. Let’s put it this way. So, manufacturing companies may use generative AI in their public relations or marketing, but those tasks have nothing intrinsically to do with manufacturing, right? And IndustryWeek is primarily concerned with managing factories and what's taking place on the floor. Those sorts of challenges. So, when I say that nobody uses generative AI, I'm really speaking to our audience and what we are concerned about. I have yet to meet anyone who has described to me an application currently in use that utilizes generative AI. I've had one person at a conference suggest a potential use case if someone wanted to design it, which sounded interesting, but also sounded like it might be more work than was worth it. I won't go into the details because I’m not really concerned with pilots in general, so I'm not really concerned with concepts either. I want to know what actually works. So that's what I mean by nobody uses it. Nobody uses it for manufacturing.
IW: It's interesting. I've seen some discussions of how it could work. We've already seen in other fields how computer programmers are using generative AI to do a lot of their code. Theoretically, a machinist could generate their own G code with some sort of AI using a verbal spoken language queue of hey, let's make that part where machining here 10% thinner on the Y axis. Something like that. It's theoretically possible, just not happening quite yet, it sounds like.
DS: You know, I'm actually going to call myself a liar and go ahead and tell you what the idea was that I heard. I think it's so funny. So, the idea was someone is walking the floor and they see a machine. Something's wrong with the machine. So, they speak out loud and say that they want to open a report, and the AI knows where the operator is standing, knows what machines are around them, and what sorts of problems are common. So the AI fills in fields or helps shape the report in such a way that the operator doesn't need to say where they are or what the machine is. The AI will kind of prompt that information and the operator can say yes or no, and the AI begins to put the report together. But the problem is, that's still real natural language processing, NLP, which is something that we already have. Generative AI is creating things fresh, right? So that doesn't really fit the definition, and that's the closest anybody's come. If I were to go back and play the tape, the subject matter expert was very convincing as they were describing the potential use case, but thinking about it later and reading the transcript, it is no, not really.
So, what's interesting about all this AI talk is that this is a language question, right? Manufacturing has been using AI for decades. Because it's a topic I know something about, I think about vision systems and machine learning training AI systems on how to detect defects on parts. It can be as simple as “does this part fit quality standards or not,” but it can also be “this part does not fit quality standards, and here is exactly what the problem is, and here's a picture of it in fine detail from any of the angles our multiple cameras can possibly see the part from.” So, we're talking about neural networks and reinforcement learning, and the technology has gotten very advanced. Once upon a time, you would have to use 100,000 images to train a neural network to be able to use image recognition in a plant and recognize which parts were bad or not. Now the technology has advanced to the point where we can only take a few pictures of good parts, and the AI is smart enough to recognize parts that don't match that pattern and tell you the part is bad. That's a huge change from hundreds of thousands of images and multiple training cycles to snap five pictures, load them into the camera, snap it on a rail, and go.
So, AI itself is something very well established. I was at a Siemens Realize Live conference two weeks ago, and in the media briefing, they were talking about generative AI, and everybody asked them about generative AI. I put my hand up and said, “Were people asking these questions prior to last year when all the generative AI hype started off?” They said, “No, not really.” And I said, “The functionality in your software that serves generative AI, was that there anyway? Was that there years ago?” They said, “Well, yes.” So, the technology's been there, but nobody was asking about it because Chat GPT wasn't a thing two years earlier. I'm generally pretty negative on hype. It kind of gets under my skin, but that's kind of where the generative AI comes from. It's got nothing to do with manufacturing. No one’s using it.
About the Podcast
Great Question: A Manufacturing Podcast offers news and information for the people who make, store and move things and those who manage and maintain the facilities where that work gets done. Manufacturers from chemical producers to automakers to machine shops can listen for critical insights into the technologies, economic conditions and best practices that can influence how to best run facilities to reach operational excellence.
Listen to another episode and subscribe on your favorite podcast app
About the Author
Robert Schoenberger
Robert Schoenberger has been writing about manufacturing technology in one form or another since the late 1990s. He began his career in newspapers in South Texas and has worked for The Clarion-Ledger in Jackson, Mississippi; The Courier-Journal in Louisville, Kentucky; and The Plain Dealer in Cleveland where he spent more than six years as the automotive reporter. In 2013, he launched Today's Motor Vehicles, a magazine focusing on design and manufacturing topics within the automotive and commercial truck worlds. He joined IndustryWeek in late 2021.