Acast sent us a fascinating piece of survey work that questioned podcast listeners about their sentiment toward AI influencing podcasts in over 20 ways. The broad takeaway is approval by listeners for AI in the production process. That approval slants toward technical improvements, with lower levels of acceptance of content manipulations.
“A new era of podcasting is fast upon us, with networks, ad tech providers, and podcasters beginning to harness the power of AI.” –Acast
The survey structure is simple, using 21 variations of this general query: “How much do you support AI being used within podcasting…” followed by prospective use cases. Four-hundred weekly U.S. podcast users were surveyed.
High rates of approval were given to technical and usability use cases. There was 82% of listener agreement with using AI to improve audio quality, and 80% approval of AI-powered language translation. Other high scores of agreement:
- Automatic transcription (79%)
- Real-time captions (79%)
- Generating reports on listener drop-off rates — more for podcasters than listeners (77%)
- Generating show notes and summaries of episodes (76%)
- Suggesting podcast episodes topically related to the listened-to episode (72%)
Only two use cases scored below 60% agreement: AI simulation of podcasters reading ads, and creation of “realistic” content, such as celebrity chat or serious debate.
Not a single use case suggested in this survey scored less than 50%; a majority of respondents approved of all suggestions.