Man vs. Machine: USC experts weigh in on ChatGPT, generative AI

March 13, 2023

ChatGPT has been the subject of controversy among AI enthusiasts and skeptics since capturing the world’s attention earlier this year. Opinions are split on its utility and competing visions for the future of assistive technology appear to clash now more than ever. Last week, USC announced the launch of the Center for Generative AI and Society to explore the transformative impact of artificial intelligence on culture, education, media and society.

USC experts are available to discuss how generative AI bots like ChatGPT add a dimension to the age-old debate of “man vs. machine” and what our expectations of the technology reveal about ourselves.

Contact: Nina Raffio, or 213-442-8464

A bridge between human and machine communication

“We are built on language as humans. We are really good at understanding subtleties, intent, and being able to communicate and organize as a society in that way. To the degree that machines are tools that we use to help us live our lives individually and collectively as a society, the natural interaction should be via language.

“Linguists don’t know everything about how our language works. We don’t really know every aspect of how we learn and computational models are a way of trying to understand.

“I think these are good tools. I do not think they are replacements for people.”

Jonathan May is a research associate professor of computer science at USC Viterbi School of Engineering’s Information Sciences Institute. His research focuses on teaching computers to understand and communicate via human language.


The problem with treating machines like humans

“What’s interesting to me is the speed and ease at which ChatGPT in particular has caused people to personify these types of AI programs. We talk about them writing and reading, we talk about what they do and do not understand, and we chat with it which is in itself a form of personification. We have a tendency to treat computer programs like they’re humans.

“What’s at stake is, if we start to think or treat ChatGPT as though it has the ability to think or reason as if it had personhood, we are going to be surprised and maybe blindsided by the kinds of errors it makes or leaves us to make.”

Mark Marino is a professor of writing at USC Dornsife College of Letters, Arts and Sciences and an expert in emerging digital literature. Read about his experiments with ChatGPT on Medium.


ChatGPT, the muse du jour

ChatGPT is not to be feared, according to experts like Patrick Crispen, who favors an “embrace and enhance” approach to the emerging technology. The approach, he says, begins with recognizing that generative AI is here to stay and comes with notable benefits.

“ChatGPT can be used to create, analyze and inspire. We’ve all been in a situation where we’ve had writer’s block. One of the ways you can get over it is by talking to someone else. Well, you can do the same thing with a generative AI platform. You can say, ‘Hey, can you give me some ideas for how to approach this?'”

But that doesn’t mean we shouldn’t be cautious, he warns.

“It’s a useful muse, so long as you realize that the muse sometimes lies. We just need to become more informed consumers of the content that it’s creating, rather than assuming its accuracy or thinking its responses are canonical. But if we can do that, we’re going to quickly realize how useful generative AI can be in ideating, outlining, writing, and editing.”

Crispen is the director of educational technology and professor of clinical medical education at Keck School of Medicine of USC. He is also an assistant professor at the USC Rossier School of Education.