‘Already here’: AI, automation find foothold in area industries | Business

Through the power of artificial intelligence, though, Mowry envisions more for Ray, the Dubuque-based credit union’s automated 24/7 digital assistant call system.

“It goes back to thinking about how a member wants to engage with us,” Mowry said. “We have wonderful service members at our locations who are always happy to answer questions, but we’ve learned now more than ever that we have to have that virtual component, too.”

Ray implements natural language processing, Mowry said, a type of artificial intelligence that helps computers better understand and use spoken language in a more human way.

If the system gets a request it doesn’t understand, the call is transferred to an operator, and the data from the call is used to teach the system how to react to that inquiry moving forward. The goal is for Ray to aid Dupaco members with minimal miscommunication, no matter the time of day.

The line is just one of Dupaco’s current and planned uses for AI, Mowry said, making the company one of many across the tri-state area and nationwide to implement artificial intelligence technology.

According to a Washington Post analysis, more than 1,000 public companies mentioned AI in their quarterly reports this summer, a nearly 2,800% increase from 10 years ago. Those companies ranged from beauty lines to airlines, financial institutions to fast food restaurants.

While the local uptake of artificial intelligence has not been as vigorous, area experts say applications of the technology locally are similarly varied, with anticipated uses, regulations and public perception expected to rise in years to come.

“For a lot of businesses, this conversation is still in the early stages,” said Jason White, vice president of business services at Greater Dubuque Development Corp. “I don’t necessarily see any new industries being created by this, but we could definitely see some changes to how existing companies operate.”

Artificial intelligence is a relatively simple term for a large — and growing — area of study. It has different definitions in different fields, and it can cover a broad range of automated tasks.

Joshua Yue, chair of University of Wisconsin-Platteville’s computer science and software engineering department, said that at its most basic level, artificial intelligence can be defined as the “computer systems that can perform tasks that would typically require human intelligence.”

But even that definition, he said, can leave out some of the nuance and history of the “multidisciplinary” field that has seen a recent uptick in public prominence.

“AI has been a branch of computer science for many years, since the middle of the 20th century, actually,” Yue said. “… But with AI being made more available to the public, particularly with (online text generators like) ChatGPT, more people are starting to think about what AI can do for them.”

Text generators such as ChatGPT and image generators such as DALL-E are examples of “generative” AI, or artificial intelligence programs that can generate some type of media. Those programs are causing most of the current buzz in public opinion and mass media, Yue said.

But there are other types of AI, too, such as artificial neural network technology and machine learning that use data and algorithms to recognize patterns and make decisions. The facial recognition program used to unlock newer iPhone models is an example of this.

Then there are all the other varieties of artificial intelligence, from predictive AI to natural language processing and automation. The list of potential applications is a long one, Yue said, and it’s only getting longer.

“AI has the potential to impact every field, to some extent,” Yue said. “It just depends on the level of automation you want and how readily available data is in that industry.”

Some well-known examples of artificial intelligence include smart assistants such as Apple’s Siri or Amazon’s Alexa, customer service chatbots and self-driving vehicles.

However, not every application of AI has to be conspicuous or public facing. For area businesses in particular, local experts say the key is not in what is most flashy, but what is most efficient.

“Businesses look at things like AI through the lens of, ‘Does this solve a problem? Does this save me time or decrease costs?’” White said. “It has to serve one or more of those needs, or it’s not worth the investment.”

To incentivize businesses to explore those potential investments, the State of Iowa in recent years has awarded over $3 million to small and mid-size manufacturers through the “Iowa Manufacturing 4.0 Program.”

The grant funding was allocated toward supporting technological and workforce investments, with an emphasis on improving efficiency. Approved uses included purchasing smart technology, cybersecurity software, augmented reality equipment and more.

“Companies have been slow to adopt some of these technologies because of the cost, … but the (Manufacturing 4.0) grant funding accelerated things and allowed companies to make investments they otherwise wouldn’t have or would have done a lot more slowly,” White said.

Dyersville, Iowa-based JAM Systems last year received $50,000 in equipment technology investment funds from the program. The family-owned company has about 20 employees and manufactures aluminum railings and fabricates vinyl fencing.

Prior to the COVID-19 pandemic, JAM co-owner Lisa Weber said, the company focused solely on packaging and distribution. But when pandemic-related supply chain issues began to affect the business, the company decided to get into manufacturing, too.

That transition was possible in large part, she said, because of the Manufacturing 4.0 funding. The money was used to purchase or upgrade equipment, including a robotic welder that has helped automate the process.

The robotic welder connects the posts in the company’s metal railings to the anchor plates that later will be attached to a deck during installation, Weber said. An employee still operates the machine, but that worker does not have to be a certified welder, as would be required otherwise.

“We knew that employees and staffing is and probably will forever be our No. 1 bottleneck in manufacturing, … so this grant came as an opportunity to streamline the process from the beginning,” Weber said, adding that the manufacturing expansion still increased the company’s overall payroll.

While AI continues to find its foothold in the tri-state area workforce and beyond, local experts acknowledge concerns still linger about its implications and impacts, particularly when it comes to generative AI.

A Forbes Advisor survey from July showed that 76% of consumers are concerned about misinformation from AI platforms such as ChatGPT and Google’s chatbot Bard, and roughly 75% expressed some concern over the potential loss of jobs resulting from AI technology.

To counter those concerns, local and national leaders have stressed the importance of regulation and education.

Federal representatives already have initiated the process of AI regulation, and last year, the White House unveiled a draft of the “AI Bill of Rights,” delineating potential principles and guidance on the equitable access and use of AI.

Some of those recommendations include proactive programming — to ensure people aren’t discriminated against via algorithms — and building in safeguards that allow people to determine how and when their data is used by platforms or AI technologies.

Educators also have had to adjust, Yue said, determining how and when to teach about artificial intelligence and how it should or should not be used in the classroom.

“AI is powerful, but it isn’t a silver bullet,” Yue stressed. “A lot of tasks and projects still need a human hand or human experience.”

White added that area employers he has spoken with have not expressed a desire to replace workers with AI. When the topic is discussed, he said, the focus is instead on finding ways to fill or relieve stress on workers from ongoing staffing shortages.

While acknowledging its ups and downs, Dubuque painter Sarah Theobald said she sees artificial intelligence as a useful part of modern technology, when used correctly.

Over the past year, Theobald has taken to using an AI image generator to create reference images for her artwork, cutting down on time taking her own photos or finding a photographer to do so.

First, she carefully writes and adjusts the prompt into the image generator until she gets a piece she is happy with. Once she has made those adjustments, she does a reverse image search of the AI-generated image to ensure it is not similar to any other existing artwork.

Then, just as before, she puts paint to canvas.

“Ultimately, art should say something about the time it was created,” said Theobald, who has a master’s degree in art history from University of Birmingham in England. “And I do believe, when done well, these (AI-assisted images) say something about our current time.”

Another field ripe for advancement in conjunction with developing artificial intelligence technology is the health care industry, where robotic-assisted surgeries have been gaining traction for decades.

Orthopedic staff at Memorial Hospital of Lafayette County in Darlington, Wis., currently are training to use ROSA, a robotic surgical assistant that assists in certain knee- and hip-replacement surgeries.

Sensors attached to the patient reflect toward ROSA’s camera, which uses the sensors’ location and a surgeon-completed calibration process to determine the exact location of the patient’s leg and joints during surgery.

Using that information, ROSA’s robotic arm holds the bone saw steady at the optimal angle while a surgeon makes the cut. The sensors keep track of any movements in the patient’s leg and adjust the robot’s arm as needed to keep the cut plane even.

“ROSA can tell me exactly where the anatomy of the bones are within less than half a millimeter,” said orthopedic surgeon Eamon Bernardoni. “That gives us highly precise accuracy in the placement of the components in surgery that can help us limit the amount of surgical dissection, … which can reduce pain and swelling afterward.”

Bernardoni, a Darlington native, joined Memorial Hospital’s orthopedic surgery team in September and has previous experience in robotic-assisted surgeries from his time at University of Wisconsin Hospitals and Clinics in Madison, Wis.

In Dubuque, Medical Associates providers also are exploring the potential advantages of artificial intelligence, said Chief Medical Officer Hendrik Schultz.

Some staff members currently are piloting an AI program that, with patients’ consent, records patient-provider conversations and analyzes the recordings to create a patient note that summarizes the visit.

The provider then can read the note and verify its accuracy before deleting the recording, Schultz said, which saves time that previously would have been used typing the note up by hand.

“A lot of time in medicine gets lost to documentation,” Schultz said. “So far, from what we have been able to see, (the AI program) is very impressive and very accurate, and that means we have more time with patients.”

Forward-looking medical applications for artificial intelligence include AI-assisted diagnosis, Schultz said, where an artificial intelligence platform would use metadata from reputable medical sources to analyze a given symptom constellation and generate potential diagnoses.

Such assistance is still a ways off for most Dubuque providers, however, and Schultz emphasized the AI diagnosis would simply be one consideration in the human practitioner’s determination of the correct diagnosis.

As with the potential for assisted diagnosis, the root of all AI programs is data. Data allows the program to recognize patterns, make predictions and recommend potential courses of action.

As more accurate, real-time data is provided to the program, the goal is to create more accurate predictions.

Financial institutions, in particular, have been using such pattern recognition programs for years to improve fraud-detection efforts and identify “anomalous” account behavior, said Todd Link, chief risk officer at Dupaco.

“If you use a credit or debit card today, no matter your financial institution, there’s a high probability there’s some AI tools being used behind the scenes to make sure that transition is as safe as possible,” Link said. “People like to ask ‘When is AI coming?’ But the answer is, for financial institutions, it’s already here.”

Link said AI is particularly good at detecting fraud because it is built to analyze mass amounts of data, taking a potentially painstaking task for a human employee and completing it in a fraction of the time.

By analyzing data from an individual’s account, Link said, a typical spending pattern emerges, making it easy for the AI to identify potentially fraudulent behavior. That behavior then is flagged for a human employee to review and follow up on.

That AI then can search through other Dupaco account data for the same anomalous behavior, Link added, allowing for early detection of any widespread breaches or scams.

“For the first time in recent history, we’ve been able to link behavioral patterns from a pool of individuals, and that’s what stops widespread fraud like what can happen if a national retailer has a (security) breach” Link said.

But the technology isn’t perfect, Link and other business leaders emphasized, and it can’t do it all.

That’s why a human employee reviews all reports of potentially fraudulent activity and remains on hand to work with Dupaco members who need assistance if their account truly was compromised.

Local experts say the same is true of many other AI applications, where human employees maintain, at minimum, a supervisory role.

“AI isn’t some monster that runs itself. It’s a tool, and it’s a tool that well-trained and qualified team members can utilize as one part of the decision-making process,” Link said.