The ongoing debate over AI and its relationship to the creative process has sparked conflict among professionals in a range of fields, from actors to designers, programmers, and now podcasters.
Now, a study from the University of Leuven, Belgium, was the first of its kind to use artificial intelligence (AI) to create podcasts about research papers submitted from the academic community.
“If podcasts could be generated by AI, that could really be a game-changer,” said Professor Philip Moons, who led the research, and expressed optimism about AI’s potential role in podcasting.
“Podcasts could be made with very little work, just by uploading the article and maybe a bit of prompting,” Moons said. “This could be a sustainable model to get the message out to people who do not typically read scientific journals.”
Moons and his team said the results of the AI generated podcasting experiments were good enough that half of the papers’ authors thought they had been created by a human, expressing shock and amazement when they were told they were listening to audio generated by AI.
To conduct his research, Moons used Google NotebookLM, an AI research assistant created by Google Labs.
“In September 2024, Google launched a new feature in NotebookLM that enables users to make AI-generated podcasts. It made me think about how it could be used by researchers and editors,” Moons said.
“When I did a first test case with one of my own articles, I was flabbergasted by the high quality and how natural it sounded,” Moons recently said, explaining his process. “At that moment, I realized that such a system could have the potential to be used by journals for science communication. Of course, the quality and accuracy needed to be evaluated. Therefore, we designed this study to evaluate its potential.”
Moons, editor-in-chief of the European Journal of Cardiovascular Nursing (EJCN) and Professor of Healthcare and Nursing Science at KU Leuven, and his team selected ten article types and contacted the authors for participation. Unknown to the authors of the chosen papers, the podcasts discussing their studies were created by AI.
These AI-generated podcasts were then sent to the authors, who commented and reviewed them for engagement, trustworthiness, and whether they could tell they were AI-made. The researchers followed up with a questionnaire created by AI using ChatGPT and a 30-minute Microsoft Teams interview.
“It was striking how accurate the podcasts were in general. Knowing that we are just at the beginning of this kind of AI-generated podcasts, the quality will become better over time, probably within the next few months. Another important aspect is that these podcasts seem most suitable for a non-technical audience, for instance, the general public or patients”, said Moons.
The academics reported that their respective AI-created podcasts captured the key points of their papers, were well structured, had a good balance in terms of length and depth, and reported that the podcast “hosts” were professional. Some of the academics listening to the podcast felt that some of the speakers or hosts convincingly sounded like they had a background in medicine or nursing, and felt the interaction between the hosts was effective.
Most authors found the information presented on the podcasts reliable, while noting issues like an overly enthusiastic tone, misused medical terms, mispronunciations, and occasional inaccuracies. The participants stressed the need for careful fact-checking before AI-generated podcasts could be released. All academics associated with the research agreed that AI-communicated content should be clearly labeled as such.
“I can see a time when all types of AI generated content could be integrated into university curriculum,” said Dr. Chance M. Glenn, President of Morningbird Space Corporation and former Professor and Provost at the University of Houston-Victoria, who envisions a time in the near future where AI-generated podcasts could potentially be integrated into university curricula in the near future to expand accessibility by offering students a variety of ways of learning.
Glenn said that the same could apply toward science journals as well, although noting that “In this case it would be important to disclose how it was created.”
“While I do believe some human touch is needed, I also believe that scientific pursuit, analysis, and dissemination is very much influenced by AI presently, and that influence will grow over time,” Glenn said. “With its capability of being the sum total of knowledge across multiple disciplines and it’s having the ability to synthesize enormous quantities of data in mere seconds, it is too tantalizing an asset to ignore.”
“At some point I can see an AI scientist pursuing certain questions on its own,” Glenn told The Debrief.
Going forward, Moons and his team’s next focus will be to discover how AI-generated podcasts can support science communication, including public and patient feedback. Moons says they are also considering the use of podcasts in summarizing scientific conference sessions.
The findings were presented at the Association of Cardiovascular Nursing and Allied Professions (ACNAP) conference in Sophia Antipolis, France, and published in the European Journal of Cardiovascular Nursing (EJCN) journal.
Chrissy Newton is a PR professional and founder of VOCAB Communications. She currently appears on The Discovery Channel and Max and hosts the Rebelliously Curious podcast, which can be found on YouTube and on all audio podcast streaming platforms. Follow her on X: @ChrissyNewton, Instagram: @BeingChrissyNewton, and chrissynewton.com.