Over the past decade, I've written numerous articles for the LSE Impact Blog and other platforms, advocating blogging as a crucial medium to share new research and ideas. Blog posts, whether as standalone pieces or companions to more formal outputs, play a significant role in the information landscape.
However, text has its limitations, and the advent of new generative AI tools has made the transition from text to speech easier. This inspired me to explore combining blogging with another passion of mine: podcasting.
Breathing New Life into Academic Articles
My goal was to revitalize my archive of written articles, many of which are still relevant and widely read. While I could record these articles as podcasts using my voice (as I possess the necessary skills and technology), I was curious about whether AI could offer a more streamlined and accessible method for academics to create podcasts.
Text-to-speech technology is advancing rapidly, with numerous tools enhancing accessibility. Audio versions are not new—audio books are well established, and publishers have experimented with audio versions of research papers, albeit with mixed results. Podcasts, still seen by many in academia as a novel way to disseminate ideas and research, have been around for two decades. According to Statista, podcast listenership in the UK is continually growing, with an estimated 21.2 million podcast listeners in 2022. So why not create "blogcasts"?
Overcoming Barriers to Podcast Adoption
The slow adoption of podcasts in academia can be attributed to barriers such as time, finances, confidence, and knowledge. However, with proper support and training, anyone can produce and share a podcast. This could be a basic, lo-fi podcast, though it still requires planning, hosting, and editing skills. Many people find recording (or even hearing) their own voice challenging. Ultimately, the content must be worthy of a listener’s time; while it doesn't have to be BBC quality, better presentation and sound quality can increase listener engagement.
The Project: Transforming Blog Posts into Podcasts
After obtaining permission from this blog and The Conversation to republish my articles in a new format, I created a new podcast account on Spotify called Talking Threads. I chose a tool called Augie, which focuses on creating videos and animations using AI. Augie allows you to add text, generating a video with an AI voice narrating it. I primarily used it to export an audio file, as the AI's choice of images was often irrelevant to the niche topics.
I introduced the podcast using my own voice and reformatted all of my old articles to improve Augie's text-reading accuracy. The AI handled name pronunciations well but struggled with compound words like "paywall" (which I changed to "pay wall") and acronyms like API and DOI (which I changed to A P I and D O I). Hearing my writing read back to me was interesting and highlighted areas needing text modifications for better audio quality. Once the issues were resolved, the process was straightforward.
Using Augie: A Brief Guide
Here's a brief guide by Augie on how to add text to generate your audio:
1. Select "create" and then choose "I have text that I want to turn into a video."
2. Paste your script, choose a voice, and preview it.
3. Once satisfied with the recording, click the three dots on the play/preview button and select "download." This will download your audio as an mp3 file.
Choosing the Right Voice
Augie offers several voices, mostly American. As a white English male, I initially chose similar voices for the podcasts. However, I realized that audiobooks are often read by people from different backgrounds than the author. I experimented with American and Australian accents and one female voice. This experience made me consider increasing the variety of voices in future projects to address representation issues.
Embracing Technology for Research Communication
My approach to adopting technology is driven by two factors: pedagogical reasons related to research communication and exploring the novelty and possibilities of new technologies. Podcasts can make ideas and written content more accessible, benefiting people with sight impairments or disabilities and those seeking a break from screen time. Portable and ambient, podcasts allow for engaging with ideas while commuting or multitasking.
While journal papers and books demand full attention, audio provides a more relaxed way to absorb information. This project underscores the potential of AI tools in transforming academic writing into engaging podcasts, making research more accessible to a broader audience.
