The everyday business world revolves around decisions. Credit approvals or denials. Stock purchases or sales. Forecasts for sun or rain.

Predictions drive those decisions, and the ability to make predictions requires the ability to see patterns.

AI is already ushering in improvements people involved in business and industry could only dream about a couple of decades ago. On the flip side, people have ongoing and even growing worries about machines taking over.

Some people are good at seeing patterns, but artificial intelligence exceeds human limits, said Alan Briggs, machine intelligence engineer for Concurrent Technologies Corp. in Savage, Md., near Baltimore/Washington International Airport.

“I think advancements in AI are going to help make those predictions better,” Briggs said. “If we can take in more data, find more complex patterns, then we can make decisions a step beyond what humans have done so far.”

However, machines don’t actually “think” in the sense of coming up with their own ideas, Briggs said.

“When we teach a machine to ‘think,’ what we are saying is, ‘Hey, can you learn from patterns of data that I’ve provided to you?’ ” he said.

‘Can I trust it?’

CTC is working to solve a few big problems lurking under the hood of advanced AI algorithms.

“I think everybody is concerned with questions, including ‘Is the output of an AI reliable? Can I trust it? Is there bias in the data? and well, what about privacy? Is my data out there secure or is someone using those?” Briggs said.

CTC is developing machine-learning AI assurance for clients in the federal government and private industry.

Machine learning is the pattern-finding, prediction-making model of AI, which Briggs believes will become the most transformative for business.

However, artificial intelligence has become popularized since the November 2022 public launch of ChatGPT by OpenAI, an American research laboratory.

ChatGPT is an example of a “generative” AI model, capable of generating new content such as conversational text, images and music from information keyed in by users.

ChatGPT is a “large language” or text-producing model.

“It’s bizarre,” said John Burley, president and CEO of diversified businesses, including ice skating rink construction company Everything Ice and Chicken Guy restaurant.

At least for small businesses such as his, the use of AI doesn’t require infrastructure beyond internet access and perhaps a paid subscription for AI models trained for specific tasks.

“The crazy thing is, I’ve very much tried to get all of our employees and friendly vendors up to speed with this because an awful lot of them still haven’t tried it yet,” Burley said. “And if you’re not, then your competitors will be. It’s kind of crazy not to do that.”

Burley uses the large language ChatGPT to clean up emails and reports – and to produce concise written policies. Another “species” of graphic AI can produce images for marketing his businesses.

He doesn’t have a graphic artist on staff and may never need one.

Burley said he can produce a photo-realistic image with a click of a button, which saddens him – despite the efficiency – because of the foreseeable devaluation of real talent, he said.

‘Unnerved’ by implications

Worries about AI beset students of Arts University Plymouth, England, where Stephanie Owens serves as dean.

“Creative fields are constantly having to incorporate innovation, but what’s different now for students is they are feeling a real challenge to their own authorship and their own originality,” Owens said. “Your training is to visualize things that other people can’t, so they are feeling very unnerved.”

There are already job listings circulating for creative people to be prompters of AI rather than being hired as film stenographers or video-game developers.

Those jobs present a dangerous path, Owens said, but she is confident that AI will become a “silent, flat note,” because AI models can only produce images based on the work of past artists.

“At some point, the problem with generative AI, in the context of creativity, is that it delivers cliches,” she said. “So if creative people want to be viable in the future, they’ve got to now trust the kind of weirdness and idiosyncrasies of their own experience.”

Owens sees intelligence as social and relational, not singular.

“I don’t think automizing a system of general intelligence is ultimately good for people or democracy because it does not allow diversity of input that drives innovation,” she said.

‘Loss of a personal touch’

A loss of individual analysis is also a danger of AI within the systems of consumer finance, said Ira Rheingold, executive director of National Association of Consumer Advocates in Washington, D.C.

His association is composed of attorneys nationwide who represent consumers who have been victims of business fraud or discrimination.

Artificial intelligence might, for example, reduce a person to a ZIP code when deciding whether that individual is credit risk, he said.

“There is a loss of a personal touch,” Rheingold said. “You are sort of just turning people into a number and making a decision based on sort of what gets spit out of an algorithm as opposed to a more individualized analysis.”

With chatbots making decisions about credit approval, the potential for bias and unlawful behavior can be exacerbated.

“If the results of credit decisions are that you know people in different classes are treated differently based on the information that is being used by AI, then companies can give the response, ‘Hey, it’s a neutral decision. We have an algorithm that decides,’ ” Rheingold said. “But the question is, what is being put into that, or what is it based on? With AI, a lot of it is garbage in, garbage out.”

AI is already at work in consumer finance and customer service. The Financial Protection Bureau and other U.S. government partners have shined an investigatory spotlight on AI chatbots.

“I think the regulators are playing catch up, to a certain extent,” Rheingold said. “I know they are looking at that implementation and asking what kind of consumer harm is happening because of it, and whether there is a discriminatory impact on decisions being made.”

‘Automation,’ acceleration

AI’s impact on workers also runs the gamut of helpful to harmful.

Fast food companies such as Starbucks use decision-making AI to order supplies.

“It’s made our life easier,” Stephen Rothhaar, a Starbucks shift manager in the Pittsburgh area, said. “On the other hand, it tends to under-order when we desperately need things.”

Almost exactly a year ago, Starbucks company officials announced the development of a new proprietary automated machine with dispensers and blenders that cut the time it takes to make a grande mocha frappuccino from about 87 seconds to 36 seconds.

The “Siren System will help drive margin expansion well into the future,” an official told Reuters in August.

Rothhaar said he would welcome the Siren System at his location, where workers have unionized for higher wages and changes in the work environment.

He hasn’t heard of any issues from his union representation about artificial intelligence.

As long as it isn’t supplanting workers, he said, he sees no problems with AI expansion.

“The company expects more and more of us, but at the end of the day, you can only go as fast as the machines go,” he said. “If artificial intelligence can help us be faster, I see that as a positive for everybody.”

About 12 years ago, the North American Hoganas metal powder plant in Somerset, Pennsylvania, added a robotic arm to take temperatures and samples of liquid steel from an arc furnace, forever changing the job of furnace operator Jerry Lee for the better, he said.

He used to have to stand within 4 or 5 feet of the molten steel to do the job, but now operators, including himself, stand at a safe distance behind an enclosed area, controlling the robotic arm.

“We are completely out of danger now,” he said.

At Hoganas’ European plants — the company is based in Sweden — Lee said automated fork trucks have been added.

“They’ve taken automation to the next step, and in our European plants, workers get moved to other technical jobs just like in my case (when the furnace operator job changed),” he said. “They sent me through a lot of training. I grew with it.”

However, as an officer of a local chapter of the United Steelworkers union, Lee has concerns about the potential future growth of so-called dark warehouses, where machines conduct full operations without human labor.

This summer, Amazon unveiled new, fully autonomous robots that carry cargo and pick up individual packages. The company plans to deploy thousands of the machines in the future, according to industry publication Retailwire.

“AI can replace a lot of jobs,” Lee said. “But AI can take over dangerous jobs and move workers to safer positions. So I’m torn on AI.”