‘Chronicle of Higher Ed’ recounts Law Library director’s challenges with AI podcasting bot

In the ways of research and academic inquiry, accuracy is paramount. So what happens when an AI tool on a popular research community website doesn’t quite have its facts straight?

A February article in the Chronicle of Higher Education recounted the experience of UW–Madison’s Bonnie Shucha, associate dean for library and information services and director of the Law Library at the UW Law School, with an AI podcasting bot on Academia.edu. Distilling her 41-page research article into a five-minute podcast, the bot made a bold implication: that she endorsed its platform. 

“I was like, Well hey, wait a minute, you’re putting words into my mouth,” Shucha told the Chronicle. 

She isn’t alone. Other scholars have also voiced concerns about the tool’s characterization of their research, according to the Chronicle. Additionally, Shucha took issue with the tool’s lack of transparency around using AI-generated voices.

“When it’s not clear [something] was AI generated, ‘that can create something that’s a little bit more scary for the listener and scary for scholarship,’” she said. 

The platform’s founder and CEO, Richard Price, told the Chronicle his engineering team are working on improvements to the tool. As Shucha noted in a February WisBlawg post, “The potential is there. The execution just needs to catch up.”

Shucha is a specialist in AI literacy and has written at length about howlegal professionals can engage with artificial intelligence tools responsibly.

Read the full story on the Chronicle of Higher Education website.