In the first installment of this two-part series, we discussed why video abstracts are becoming a popular route for scholarly article authors to take, as a way to increase both quality and visibility for their submissions. Video abstracts can provide a brief but clear summary of an article, with attractive and attention-grabbing visuals designed to hook readers as they scroll through an online page.
In some cases, though, the brevity of a video abstract can be a hindrance. Certain types of scholarly articles simply cannot be summarized in the short time period that a video abstract provides. Moreover, for some types of scholarly research, audio is more effective than video.
…So, this is where another multimedia form comes in—the podcast!
Podcasts are defined as episodes of audio-based series that can be downloaded or streamed. They can be used in lieu of a written article or as an addendum to one. Either way, though, as is the case with video abstracts, authors should ask themselves a series of key questions before getting to work producing a podcast as part of an article, or as the article in its entirety.
The first step is determining whether to go the route of a written article with a podcast attached, or to use the podcast in place of the written material. Once that decision is made, it’s time to decide between a podcast in an interview form, or one that features a single voice. And finally, since literally everything in today’s publishing world involves some degree of AI usage, authors must tackle the big question on everyone’s mind: “How much can AI help me create my podcast, and what still needs old-fashioned human help?”
Instead of, Versus in Addition To?
Podcasts can be submitted both on their own as complete articles, or as an audio file to run with a written piece. But how should authors decide on the best route?
The first step, of course, is to see what guidelines (if any) exist for the particular publication to which the work is being submitted. Because podcasts as part of—or in place of—scholarly works are still so new, each publisher is going to have very different policies about what is, and is not, permitted. In addition, if standalone podcasts are allowed, the publisher likely will have certain stipulations, such as a word limit for the transcript or mandatory mentioning of keywords (much like would be the case for a written article).
Beyond that, though, another big question is: How important is the human voice component of the article?
Take, for instance, a submission to a medical journal that deals with a patient describing daily life while coping with a chronic illness. The element of emotion is critical for this type of subject matter, and the author may not be able to capture that emotion in written form. So, in this case, a standalone podcast might be the best option.
If, on the other hand, the emphasis of the article is on the biology of why the illness occurs—as opposed to what daily life is like with the illness—readers might find a written format easier to follow, due to the complex nature of the points the authors are trying to make. Although it’s possible for readers to stop and start a podcast if they did not understand something the first time, many people might find it much easier to read a written version more than once. That being said, though, a patient testimonial about life with the illness still might provide value to the article in the form of a podcast as an addendum to the written piece; it just might not be as effective for the podcast to completely replace the written work.
Granted, there is no “right” or “wrong” answer to this question, and it’s a judgment call (unless, again, the submission guidelines state otherwise). As time goes on and more data exists, it will likely be easier for authors to make this crucial decision.
Many Voices or Just One?
There’s no “right” or “wrong” answer as to whether to make a podcast a back-and-forth conversation/interview between two people, or if one voice is enough. It all depends on several different factors.
The first question is whether one person’s overall tone and diction is strong and sufficient enough to carry the entire podcast solo. A science researcher with a background in public speaking, for instance, might be able to make this work. On the other hand, someone whose background lies more in laboratory experiments may need more help, since they have less experience with similar types of delivery.
Another aspect to consider is the content and subject of the article. For a medical article that covers both patients and providers of a particular disease or condition, an interview between those two parties would likely be the most effective way to go. But if the article instead deals with a breakthrough discovery that one person made, it might be more eye-opening to allow that one person to take the lead for the entirety of the broadcast.
And, of course, there’s always budgetary considerations. If each person participating requires a fee, there may or may not be monetary room for more than a single voice. This should be taken into account before beginning work on the podcast. After the fact, it could make production much more complicated than anticipated, or it could necessitate major changes that slow the entire process down—not ideal, especially if deadlines are tight.
The Right (and Wrong) Way to Use AI
In part I, we discussed the ways in which AI should—and should not—be used in the process of creating video abstracts. Appropriate times to use AI might be, for instance, during transcript and caption generation.
The same is true for scholarly podcast creation, and there are even numerous tools available that are specifically designed to convert audio into a podcast—NotebookLM, for example. These tools create AI-generated voices to bring a casual-sounding tone to that audio, making it easier for listeners to follow.
But authors must be careful, particularly if their content requires permission. While permission credit lines might be easy to add to written work, for audio, it might not be that simple—and the end result could be rejection for plagiarism.
Moreover, AI will, at times, inevitably mistranslate raw audio. This means authors must carefully check any AI-generated audio material for inaccuracies. Failure to do so could mean a factually flawed finished product, necessitating time-consuming corrigendums and potentially damaging a journal’s reputation.
…The bottom line? AI never has been a perfect tool and never will be—including, but not limited to, when used for scholarly podcasts.
Possible Directions for Future Research
The use of podcasts in scholarly writing is still a brand new topic within the publishing field. And as is the case with any new topic, more research is needed to nail down the efficacy of this medium.
In the coming years, academic professionals should conduct surveys among both medical and science students—whether at the undergraduate or postgraduate level—about whether they feel articles in podcast form are interesting or helpful. The potential advantages to podcasts over written articles are abundant, such as ease of access and the ability to quickly transport audio files to different places. Still, there are possible complications, such as the inherent risk of becoming distracted by trying to listen to those audio files while multitasking—an issue that is more difficult to encounter when reading an article in written form. It will take many years of data from surveys to determine if the benefits outweigh the risks.
By Anne Brenner
