WASHINGTON — The adoption of artificial intelligence (AI) is increasing rapidly across industries, and manufacturing is no exception. 

While most manufacturers report being in the early stages of AI adoption, 78% plan to increase AI spending in the next two years, while 52% have a corporate AI strategy, according to a survey conducted by the Manufacturing Leadership Council (MLC), a division of the National Association of Manufacturers (NAM).

At Nexus 2024, held Sept. 30-Oct. 3 in Washington, DC, Penelope Brown, senior content director, MLC and NAM, detailed the benefits AI can offer manufacturers, misconceptions around the technology and how to implement it effectively.

“AI is as good as it’s ever been and as bad as it will ever be,” Brown said. “We’re kind of at the starting point with all of this right now.” 

For baking and snack manufacturers, AI can help optimize product portfolios, Brown said, analyzing SKUs, product sizes, packaging and more to determine which products are worthwhile investments or could be discontinued. 

AI-guided vision systems can also improve quality control, Brown said, citing a company that adopted this technology.

“The problem with the standard vision system is it usually is pretty limited in what it can detect,” Brown explained. “But [the manufacturer] was able to implement some deep learning on these AI-guided vision systems to find areas that really weren't detectable before, so things like twisted cables or scratches off the surface or cracks.”

Analysis of production efficiency, as well as predictive maintenance assistance, are two other growing AI applications in manufacturing. 

“Using an [AI] algorithm to determine if production is on schedule, for example, or to identify a bottleneck that might need to be impressed, seeing if your supplies are going to be delivered on time,” Brown said. “Predictive maintenance is a really big one that we’re seeing, basically raising an alert when a machine operates a little bit out of spec.”

Despite AI’s growing capabilities, Brown noted there remain misconceptions about the technology One is that it will decrease jobs in manufacturing. However, Brown noted this is unlikely given the current labor shortages across manufacturing.

“As of July 1, there were 486,000 unfilled job openings in manufacturing,” Brown said. “So right now, the industry can't hire for the jobs they have. It's very unlikely that AI is going to replace all of that anytime soon.”

Another misconception is that AI is too expensive and complex to be adopted by smaller companies. However, Brown observed that tools such as Open-Source AI, which lets users freely access, develop, modify and share AI technology, is allowing manufacturers to adopt AI at a lower cost.

For manufacturers adopting AI, Brown recommended they have a clear understanding of the data they’re collecting, defined objectives and a business case for AI adoption, a dedicated individual or team in charge of AI, and buy-in from anyone who will be using this new technology. 

“Talk to your frontline workers,” Brown suggested. “They're going to tell you, ‘I've got an issue with this thing, it breaks down all the time. This never comes out right.’ That's a really good place where you can find the most effective use cases [for AI] and really have the most impact on the operations.”