Opportunity and concern come with AI, CSUMB experts say

By Mark Muckenfuss

From mental health screening to assuring financial compliance, streamlining lessons and even reading minds, Cal State Monterey Bay professors are using artificial intelligence in ways that utilize its efficiencies and stretch the possibilities of this new frontier in technology. 

“My research has to do with brain data,” said Glenn Bruns, an associate professor in the School of Computing and Design. “We’re collecting EEG data and doing what you might call mind reading. We can use deep learning to guess what you're thinking.”

The project, Bruns said, came from the idea of being able to find previously read web pages just by thinking about them. It was further spurred by an acquaintance’s struggle with amyotrophic lateral sclerosis, also known as Lou Gehrig's Disease or ALS. The disease inhibited the man’s ability to speak. 

Bruns postulated that if you could guess a concept that a person was remembering, based upon brain activity in a particular area, it might be possible to guess a word a person was thinking. 

“You could predict what they were saying,” he said. 

So far, his research has only involved a single subject. Bruns said the predictions were better than random guessing, but “not great.” He’s planning to do a larger study. 

It’s one example, he said, of how AI can make possible what perhaps only a few years ago, seemed more fiction than science. 

“We don’t even know where this is going,” Bruns said of the technology systems that have been on an exponential growth curve and are slipping into many aspects of our lives. “It’s not totally understood how they do everything they do, but I think the abilities of AI systems will increase dramatically. I see cures to a lot of diseases being solved by AI.”

Recently, he said, he used ChatGPT, perhaps the AI application most people are familiar with, to give him a summary of the Netherlands’ World War II history while he was traveling in the country. 

“It’s like a super advanced web search,” he said. “You just say in English what you want and the AI is using everything it's seen -- billions of documents -- and giving you a cohesive summary.”

Jill Hosmer-Jolley says she often uses ChatGPT to help her with her course materials and lectures. 

“Last night, I uploaded assignments and said, ‘Can you make these more concise?’” Hosmer-Jolley said. “And it did a beautiful job of cleaning it up.” 

Hosmer-Jolley is a lecturer in the College of Business and a proponent of using AI in the classroom. She recently attended the Ai4 2024 conference in Las Vegas, where she was the moderator for a panel on integrating generative AI into workflows and operations. 

While some of her colleagues shied away from the early versions of generative AI, such as ChatGPT, Hosmer-Jolley embraced them.

“I fell in love with it,” she said. “It's really about employability with me. If someone hires a college grad, they want them to know the most current stuff. I started running workshops and putting it into all my assignments.”

While some fear students will use the tool to avoid work, Hosmer-Jolley said she uses the process of incorporating AI into assignments as a way to get her students thinking.  

“I ask them to open Google Docs or Word and initially answer the assignment orally,” she said. “Then they can put it into ChatGPT and say, ‘Clean this up for me.’ They can develop it from there. They have to copy the entire chat text and put it at the end of the paper. I get to see how they developed their thought process.”

Some worry that that very process may be lost as AI takes over more and more functions. 

“I worry about people blindly using it and not having the education about when it’s useful and when it’s dangerous,” said Sam Ogden, an assistant professor in the School of Computing and Design. As people surrender more of their critical thinking to AI, he said, “It could cause us to be more and more average and less exceptional. That’s my biggest fear.”

But he also sees great benefits. Ogden has been working on using an AI algorithm to screen for when people might be experiencing mental health problems.

I’ve been doing a lot of work with one of my collaborators on doing mental health screening,” said Ogden, whose background includes outreach fora student counseling center during his graduate studies. “This is a really cool thing we can do.

The team correlated phone message response times with anxiety and depression, discovering that a delay in normal response times often indicated mental stress. 

“We found we could make these really strong predictions,” Ogden said. “If you’re seeing suddenly that someone is taking a long time to respond, check in on them. They might want to talk to someone.”

To put the attention on AI in perspective, consider the turnout at the International Conference on Machine Learning in Vienna this summer. Ogden attended and said there were 2,900 presenters. Hosmer-Jolley said an average of 350 AI applications comes online every week. 

“I’m picking up (a class) in advance machine learning this semester,” Ogden said. “I’m both terrified and excited to pick it up because the field changes so fast. It’s utterly overwhelming but fascinating. I’m going to have use AI to stay ahead of them.” 

In the business arena, Gary Schneider said AI is making the work of accountants and auditors easier. Schneider, an accounting professor, said he is interested in seeing how AI will “change the world.”

In the realm of accounting, he said, AI is already saving auditors hours of examining lengthy financial and legal documents.

“Accounting firms use AI to read those documents, identifying potential exceptions that human auditors can choose to examine further,” Schneider said. “It might look for things that seem out of whack, or it might look for inconsistencies within the document or between it and other client records.”

But there are limitations.

“Generative AI can do some pretty neat stuff,” he said, “but it isn’t that smart yet. If you ask ChatGPT, ‘What’s the population of Turkey?’ it will give you an answer. But it won’t necessarily  be able to tell you where or how it got that answer.”

The next step in the technology is making that information available, he said. 

“Researchers are now working hard to develop explainable AI,” Schneider said. “The major accounting firms that are building their own AI tools have all sunk a lot of money into this.”

“I think there is a degree of FOMO,” he added, using the acronym for fear of missing out, “companies saying, ‘We can’t afford not to chase this.’ A lot of this is defensive strategy, but it’s also, ‘Hey, we might stumble onto something valuable,’ too.’”

Whatever the results, he said, as AI improves, it will radically change things, from reducing the time it takes to make vaccines to improving the editing process in music and video production. It will also impact the workplace in ways that will only be revealed over time.