When we hear the words artificial intelligence, a lot of different emotions and opinions come to mind.

Machines making decisions on their own, taking away jobs from humans is often the thought our minds come to, but at least in the manufacturing world, that is not the goal, according to Ralph Hardt, president and owner of Belleville International in Butler

Hardt believes the phrase artificial intelligence is very generic because there are so many different ways the technology can be applied.

“I think the phrase AI is overused sometimes,” Hardt said. “It should be functional AI versus theoretical or destructive AI. There is this functional AI world and this theoretical AI world, and I think they are two different things.”

Belleville International, located at 330 E. Cunningham St., manufactures specialty-style disc springs, spring washers, flange bolting, valves, and other precision-machined and stamped components.

Just a few weeks ago, Hardt said Belleville implemented a form of “functional AI” with a Zeiss Coordinate Measuring Machine.

“It involves a compensation software,” Hardt said. “It involves a robot as well. It will allow us to run more unattended machine time, which means we will get more hours out of our machines. It will result in faster lead times for our customers and result in improved quality.”

The machine, which cost nearly $1 million, has yet to be used, but Hardt says when it is, it will take data from the measurements of parts it is making and will calculate the offsets it needs to make in order to produce the finished product.

“That process would normally require the machinist doing the calculations to shut down the machine, typing in the tool offsets and then restarting the machine,” Hardt said. “The robot takes the part off the machine, measures the part, and then, through the software, calculates the offsets the machine needs to do.”

In the meantime, Hardt said Belleville is “babysitting” the machine until they completely understand how it functions. Once running, it will work continuously.

This is just one aspect of how AI can be utilized in the manufacturing industry, but according to Shah Limon, associate professor in the department of engineering at Slippery Rock University, there is much more.

“In a manufacturing process this technology helps to understand your machines,” Limon said. “It allows you to predict things ahead of time so you can make decisions based on that. The market is very competitive. So if you can use the AI to make decisions which helps you understand the future it’s is a big advantage.”

Limon said another example of AI being implemented across manufacturing is the maintenance of its systems and machines.

Instead of waiting until something breaks, or replacing parts before they need to be replaced, there is software capable of detecting such things in order to improve the life cycle of parts.

“The problem with routine maintenance is in one month that machine might have only been used for 100 hours,” Limon said. “The next month maybe 200 hours. What the AI technology can do is with a sensor it can monitor the equipment constantly. With the sensors it can predict when it will fail.”

According to Govardhan Singh, site head of Bayer’s Saxonburg facility at 150 Victory Road, the company is actively looking into implementing a form of this AI into its manufacturing process.

“What we want to do is take machine data and feed it into AI to see if we can predict breakdowns of machines,” Singh said. “This is an area we want to start dabbling with.”

Limon said AI also can be used to check the quality of finished products in manufacturing, which again used to be a manual process that oftentimes led to inconsistent results.

“With a camera it takes pictures which goes to the cloud computers and with an algorithm it can make sure the product contains the quality you want,” Limon said.

According to Abhinav Srivastava, director of engineering at Bayer, the company’s Saxonburg facility already has implemented a form of generative AI for quality checks on its final product, particularly in its radiology products.

Generative AI is capable of generating text, images, videos, or other data using generative models, often in response to prompts.

“There is a lot of buzz on generative AI,” Srivastava said. “We have condensed it to Bayer’s use. It does not go out and get data from across the internet. It’s limited to what Bayer can do for its product.”

Bayer puts customer complaints into its AI software, called myGenAssist, which in turn can identify why these problems brought up by customers are occurring.

“For example, say we take 100 complaint items, put them in the prompt and then ask the generative AI what are the top two issues,” Srivastava said. “Then it will go through each line item, and then it will tell me the top two items. Prompt engineering is the heart of this generative AI.”

Srivastava said Bayer has been using myGenAssist since the summer of 2023, and at this point, the machine is now able to identify issues before there is a strong trend of customer complaints to begin with due to the fact it has more data to pull from to identify issues.

The cost of machines and systems such as these can prove to be difficult for some smaller to medium-size companies, Limon said.

It can also just depend on the nature of each business, as it might make sense for someone, but not others.

“Suppose you have developed a new product and you don’t know how much demand it has,” Limon said. “Then it might not be a good idea to implement the AI. If the company knows the product is secure and has higher demand, then the cost of the AI would be justified.”

Limon said another cost companies should consider after implementing AI is upgrading their cybersecurity systems, as some of the product data or other information will be uploaded to cloud servers, which can be prone to cyberattacks.

The big question on most minds is will AI take away jobs from humans. According to Limon, the answer is both yes and no.

“There are jobs that will be replaced,” Limon said. “That is an undeniable fact. This has been happening for hundreds of years. Some jobs will be gone, but at the same time common jobs will be created because the technology creates new opportunities.”

Hardt had a similar response, saying he believes jobs will be shifted instead of eliminated.

“They will be managing a process versus typing into the machine control,” Hardt said. “It still requires a skilled machinist to run it all. At Belleville we have seen continued job growth.”

Hardt said Belleville’s product production has grown by 33% over the past year, including its export of products to 14 countries, which makes up 25% of its sales.

“People don’t want to just load and unload parts anymore,” Hardt said. “The younger people grew up with this technology, and we will bring in more as it makes sense.”

While technology has been more readily available to younger generations, there still needs to be proper training and education to fully understand it.

Limon said all SRU engineering students have to take programming classes, which helps them understand the coding and mathematics behind the technology.

“For industrial engineering, I teach AI classes,” Limon said. “It talks about changes coming. We also have students go out into the world and do real-life projects with Western Pa. companies. We also talk to companies and ask what changes they have coming up so we can put them into our curriculum.”

Hardt believes more companies are using AI in some fashion, which in turn will help production, thus fuel expansion.

“Those that don’t stay ahead of it will become less competitive,” Hardt said. “We all know what happens to companies that become less competitive.”