Leveraging AI in the workplace and education brings both pros and cons

As artificial intelligence becomes more closely entwined with everyday business, students with a deep understanding of generative AI will be in higher demand when it comes time for employment.

The future with AI is creating anxiety for both universities and industry, but those who embrace the inevitable expansion of technology will be better prepared for success.

“I think employees can look at it as an opportunity to leverage these tools to make themselves more valuable in the marketplace,” said Michael Mior, assistant professor in the department of computer science at Rochester Institute of Technology’s Golisano College of Computing and Information Sciences. “One thing that our department has heard from our industrial advisory board is that they’re really looking for graduates (in the computer science field) who are able to make use of AI tools, people who are comfortable using AI tools to help them program.”

Clearly there’s no hiding from AI. It’s here to stay, and its uses are only going to expand exponentially.

“Our students are embracing it and using it,” said Maria Richart, director of career services and cooperative education at RIT. “It’s going to change the way we work. But we’re also bracing for it to be a positive change. That’s what my employer partners keep telling me.”

Businesses are exploring uses and determining how to incorporate AI systems into everyday life. At some point, it will surely mean a reduction in certain areas of the workforce as companies look to increase profit margins.

“A lot of these tools are not at the point where they can replace human workers completely,” Mior said. “I think it’s true for a lot of businesses, if these tools are adopted appropriately, they might be able to reduce their head count in some areas, but I don’t think it’s largely the case where you’re going to have entire departments eliminated, at least not at this point.”

One reason: generative AI is not infallible.

“It doesn’t know what’s false, what’s right, it just knows given what you’ve asked it to do or what you’ve written, what’s the next word it should say,” said Chris Kanan, associate professor of computer science at the University of Rochester’s Center for Visual Science, Brain & Cognitive Sciences.

Indeed, there already have been plenty of instances where AI tools were used as a convenience and the results were a colossal failure. Such as at the Bay Area auto dealership in California, which was using a chatbot on its website.

“A customer had a conversation with this chatbot and then convinced the chatbot to say, ‘I’m going to make you a legally binding offer to purchase a vehicle for $1,’” Mior said. “If you had a human employee doing that, they’re potentially losing their job, they’re potentially legally liable for their actions. When you have this AI chatbot doing this, what are your repercussions there?”

There’s also the story of a lawyer using ChatGPT to write a legal brief, which was then submitted to the court.

“It turns out that some of the cases that were cited were completely made up, they didn’t even exist,” Mior said. “That was ultimately the lawyer’s fault for not even confirming the output of this tool.”

Which brings us to one problem with generative AI: it can’t learn on the fly.

“They make stuff up, they can’t learn continually,” Kanan said. “The creativity is a pro and a con. It doesn’t know when it should be creative and when it should make stuff up, and it has no ability to tell the difference between the two.”

But there are many uses for AI that don’t require creativity, or even allow it, and that’s where implementation will be used soon.

“There’s a lot of jobs that are very menial,” Kanan said. “Even hospitals like URMC employ a significant number of nurses that just need medical training to take unstructured data, turn it into structured data and enter it into databases. You can do that with (AI), no problem.”

Said Mior: “The things that AI is good at in general is pattern-matching. Whenever there is a lot of repetitive work, whether it’s data entry where we have the same task being performed over and over again, those are the sorts of things AI can end up doing very well, generally speaking.”

Other things AI can do well: homework. So as schools ensure they are providing every educational opportunity necessary to prepare students for employment in a world of artificial intelligence, some in academia worry about how the tools can be misused.

It’s a bit of a conundrum. Employers want the students trained in the latest technology, but what are the educational consequences?

“You can look at it as a way for students to cheat the system and not do their own work,” Mior said. “But then there is also a way to look it: how can we actually integrate this into the curriculum and get students comfortable using these tools.”

Kanan said that’s really the only way to approach AI in the classroom.

“The students who wanted to could always cheat,” Kanan said. “If people want to cheat, now everyone can cheat. On the other hand, I feel like the students have to have a responsibility to want to better themselves and learn.

“Future students are going to grow up with these tools, and they’re going to be expected to use them in the workforce. If you try to tell them this isn’t ever allowed, they’re not going to respect the education.”

[email protected]/(585) 653-4020