‘Sorting signal from noise’: Academic podcasting in an age of excess

SETU’s Rob O’Connor talks academic podcasting, disinformation and the ethics of AI.

Rob O’Connor is a computer science lecturer and researcher at South East Technological University (SETU). His current research focuses on the intersection of podcasting, scholarly communication and artificial intelligence (AI).

A Waterford man, he’s also a musician and former radio DJ. He used to host an Irish music show where he interviewed artists. “For a period there, I think I knew everyone in the Irish music scene because I spoke to them all,” he tells me.

The way he tells it, he sort of fell into computers – he got the points in his Leaving Cert for the applied computing degree in what was Waterford Institute of Technology (WIT) (now SETU) and ended up loving it. “I really took to the programming,” he says. “And I can find my way around any problem if I have enough time.”

After working in England and Australia for a couple of years – doing “all these typical things that Irish people do” – he came home just around the time the dot-com bubble burst and there were very few tech jobs in Ireland. This is how he ended up back in Waterford completing an MSc before taking up a full-time lecturing job at WIT.

The MSc was in network management and he recalls heading up to the Young Scientist and Technology Exhibition around 2002 and doing a Wi-Fi demonstration for the students. “It’s hard for anyone under 30 to really understand what the world was like before Wi-Fi,” he says. He laughs at the memory of taking a photo with the education minister at the time, Noel Dempsey, to celebrate this “internet without wires”. It’s twee when you think about it now, O’Connor says, but it just shows how much things have changed in the last 20 years.

In 2017, his wife Dr Jenny O’Connor, who lectures in English and communications at SETU, wanted to start a podcast as a way for her students to engage with the coursework in a different way. She hosts ‘The Nerve’ and he handles the technical side of things. The following year, he decided to follow her lead and do a podcast for his own students, ‘The Machine’. He has since also worked on ‘9plus’, a SETU podcast that highlights research being done across the university.

Why podcasting?

O’Connor thinks podcasts are a great medium to make complex topics more accessible and engaging because they force you to adopt a more conversational tone. “You can take a topic and break it down into a series of steps or explain things through analogy or metaphor to help people understand.”

He says the Covid-19 pandemic highlighted the need for accessible science communication with more people than ever looking to understand complex scientific processes, and more disinformation than ever to wade through to find the facts.

One thing podcasts can be really useful for is communicating uncertainty, O’Connor says. He talks about how people who spread disinformation can often speak with confidence and charisma, whereas a scientist is much more likely to speak in terms of sober data and probabilities. “I’m not saying everyone needs to be a performer with jazz hands,” O’Connor emphasises, “but we need different ways of speaking with people and providing high-quality information.”

Podcasting isn’t the only way to engage wider audiences, he says, but it’s one way he thinks hasn’t yet been fully exploited by academia.

‘7 Ps of podcasting’

Around this time of Covid disinformation, O’Connor’s sister in law asked him if he thought about doing a PhD about scholarly podcasting and it was music to his hears. He gave up the radio side gig and registered for a part-time PhD at Adapt Research Ireland Centre for AI-Driven Digital Content Technology in Trinity College Dublin. O’Connor has known his supervisor, Prof Owen Conlan, for many years, and Conlan’s own research looks at empowering users in understanding and interacting with complex information and media, so it sounds like they make a good team.

One of the things O’Connor hopes to achieve with the PhD is to develop a framework for academic podcasting. Many individual researchers and institutions are developing podcasts but there’s no formalised approach or repository of tips or tools for doing it.

“It’s a different skill, a different way of communicating,” O’Connor says. After years on radio, he knows when to ask the more basic questions or when an explainer might be necessary and other ways to keep an audience engaged and informed. As well as that there are lots of technical aspects to recording, editing and publishing a podcast that he’d like to develop standards for to support academics.

He’s currently working on his ‘7 Ps of podcasting’ – “you need a snappy title” – planning, preproduction, production, post production, publication, promotion and preservation – with guidelines and tools under each heading.

Preservation, in particular, is an interesting topic. “At the moment, there’s very little preservation,” he says.

Part of the argument for creating a podcast is to make research more accessible and so you have to think about where that podcast is going to be hosted during your research, but also what will happen to it after your project has ended, and how the platform that hosts it will use your data.

Suppose you host your podcast on a particular platform, what if it shuts down tomorrow or if it changes its terms and conditions, he says. These are issues O’Connor is thinking about.

He thinks an EU repository would be a good way to create an open-access ecosystem for podcasting. “I would love down the road that there was a European platform … a space where open-access podcast materials can be made available, and they live beyond the research project.”

The EU is “a suitably weighty non-commercial entity” to work with so researchers could avoid some of the ethical and practical challenges of working with tech platforms.

Ethical considerations

It’s not long before we get talking about AI. For his PhD, O’Connor is think about automation more so than large language models, so looking at what tools are available to help with the tech side of podcasting. However, as a lecturer he has to think about generative AI (GenAI) and its implications for teaching and learning.

He was a bit surprised to find that plenty of his students have ethical concerns about using AI, from how their data might be used to the environmental cost of its energy use. He says there’s a misconception that all computer scientists are “tech bros” who buy into the industry hype. “We’re not all like that,” he says.

He’s not totally against GenAI either though. We get to talking about a concept that informs his thinking about the value of scholarly podcasting – ‘the crisis of complexity’, that is, the challenge people face when trying to stay up to date on a subject because of how much scientific knowledge is being generated all the time.

The medical field has really leaned into podcasts, he says, because of this challenge. When you have an expert audience with a particular need, such as with medical professionals, O’Connor sees this as a use case for AI-generated podcasts to allow people to keep on top of the latest research and techniques.

GenAI “can be a great productivity enabler” when you already have in-depth knowledge of a subject, he thinks. He gives the example of coding – if he uses an AI model to generate code, he’s able to spot the errors quickly, whereas if you don’t have that knowledge, it’s especially tricky because it can give “a confident answer that looks plausible” so it’s harder for a novice to spot errors.

It’s a complex issue, he says. “And, truthfully, I don’t think anyone has the answer and maybe there is no one answer.”

Thanks for listening

As for the value of podcasting, he’s more certain about that. I ask how his students have responded to the podcast. Anecdotally, he says they seem to really engage with it. He recalls how one student shouted across at him on campus one day – “when is the next episode out?” He loved that.

He gets the students involved in making podcasts as well as part of their coursework and he finds they’re always really interested. It’s also a great skill to have.

When companies come in, they always say they want the computer science graduates to develop their soft skills. And working on a podcast in interdisciplinary teams really helps build up those skills and that will stand to students, he thinks. “You can have brilliant ideas but if they’re locked in your head, they’re useless.”

At the core of O’Connor’s work is a desire to get more evidence-based science into the public domain.

“If you think about the crisis of complexity, there’s so much stuff out there, it’s very hard to distinguish the signal from the noise.

“And that’s what I’m talking about [with academic podcasting].

“It’s not about creating more noise, it’s trying to help people create more useful signal.”

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.