7 Comments

I agree that in the most part, human is much better. However, I can also see a role for the generated audio (esp. as the accuracy improves) for students to use independently - in the same way as text to speech may be preferable to reading.

Expand full comment

Thank you for the comment. I have concerns about comparing text read aloud to AI generated podcasts. Read-aloud allows students to experience authetic text written by a human - actual information. An AI-generated podcast is slop with no intent and likely accuracy issues as I documented. I address the idea that generative AI will inevitably improve here (https://www.criticalinkling.com/i/148497801/todays-ai-models-are-the-most-rudimentary-ones-our-students-will-ever-use).

Expand full comment

I agree from the content point of view, it was more thinking about the voice aspect that I'd used that comparison.

I'd also been thinking about a student who'd told me she'd read an article, then listened to the podcast version in the bath - to help her remember the key points. Her comment was that it was generally useful though there were a couple of points she thought it should have included & didn't. I agree that she wouldn't have known about those points had she not read it first.

Expand full comment

Thanks Tom! I would not trust an AI podcast and you demonstrate why. Especially not for students.

Expand full comment

Thank you, Janet! Do you mean to tell me you're not looking into flights to San Francisco to visit the city's famous leaks? 🤣🌉🚿

Expand full comment

Crazy! I don't want to speak with someone else's words - someone's stolen words. Why would we speak to students with stolen words - then expect that they won't plagiarize?

If someone has a speech - related disability maybe the technology is appropriate. Otherwise let me hear your own voice!

Expand full comment

I think speech-to-text which conveys actual information is much more appropriate.

Expand full comment