AI Podcasts and the Echo of Authenticity

Google’s Illuminate AI podcast feature received a series of major updates that eclipsed NotebookLM’s Audio Overview feature, which first made waves in education. Illuminate isn’t the only AI podcast update that has made recent news. HeyGen, the AI avatar company, also shipped an avatar podcast feature. Now that we’re awash with AI podcast apps, has anyone had the time to ask what assigning synthetic voices to our students means to the authentic relationships we try to form within our classrooms?

Illuminate’s Major Updates

I’ve written about Illuminate’s podcast features before and this latest round of updates allows users more access and control over the generated output. Illuminate is built around AI audio, so their is a greater focus on the quality and control v. NotebookLM. Illuminate also includes a transcript of the AI podcast—something NotebookLM hasn’t gotten around to shipping yet. Previously, the app was locked to open-access articles found on specific research databases like arXiv and other computer science-related sources. Now, you can upload multiple sources from across the web.

The one major knock many people have against the Audio Overview feature within NotebookLM is the lack of customization and user control. Illuminate now allows you to pick a variety of styles from six voices. This level of customization gives the experience a much more fluid listening experience compared to the default voices within NotebookLM.

Stylizing the voices via prompt instructions is also another option that doesn’t exist within NotebookLM. You can make the AI voices talk like pirate or modify them further to sound like just about anything. You can also listen to an example to hear a recent Edsurge article I contributed to an interview.

HeyGen’s Video Podcast

AI Audio podcasts got some competition from HeyGen with the newly announced Video Podcast feature. Now you can watch synthetic hosts awkwardly jerk their bodies and perform an AI summary of a URL. The results are obviously lackluster compared to Illuminate or NotebookLM, but you’d be foolish to dismiss it outright. We haven’t hit the ceiling in what these models can be tweaked to do.

If you’ve followed generative AI deployments for a while, you’ve noticed this latest AI podcast explosion follows a familiar trend. An app or use case gains traction on social media and gets copied and incorporated into competing apps. The difference with genAI is it isn’t just apps that join the bandwagon—it’s also modalities. We’ve barely had time to consider the implications of synthetic voices entering our classrooms, now we’re seeing the same podcast feature appear in digital avatars.

Education is just one field that has to contend with this emerging phenomenon of rapid AI deployments and updates. Private industry is in a similar boat as every teacher right now, left to wonder how these new features will impact their messaging, their brand, and their product. What sets education apart, though, is many of us are worried about what these innovations will mean for the relationship we build with our students.

Real Presence and Authenticity in an Age of Rapid Transformation

I had the pleasure of attending Liz Norell’s launch for her first book The Present Professor: Authenticity and Transformational Teaching. It is a terrific book for our time, and I don’t just mean dealing with AI disruption. Many of us feel disconnected from our teaching and our students, but as Liz argues, it may well be the teaching persona we spend so much time developing that creates unneeded barriers between our authentic selves and our students.

When you take the time to examine your own presence in the classroom and peel back the layers that cause you stress, then you can unlock more meaningful ways to improve your sense of self. Doing so isn’t a cure-all for what ails us or complicates our teaching, but I think we all agree that students respond best to authenticity.

It took me years to realize I didn’t want to be in front of a classroom wearing the persona of what I thought a successful teacher should be. That created barriers between how I communicated with my students, and likely how students responded to me.

The big question I have about these new multimodal AI tools is what deploying them will do to the relationships we value. A student turning to an artificial voice or avatar to gain understanding isn’t damaging on its own. We can engage nuance and admit that these tools have promise to improve certain educational outcomes of our students. But that’s in an idealized world, one where developers work hand-in-hand with teachers and students, instead of releasing a technological hurricane every few weeks. The reality is most of us haven’t had the time or resources to engage AI.

We Need Better Frameworks for Dealing with AI—Not Just Adoption or Resistance

We talk about safety and existential threats when we talk about AI, but rarely do we consider the impact the uncritical adoption of this technology will have on authenticity and presence within the classroom. Such synthetic cosplay may well one day pass as a replacement for the genuine care and relationships we build within ourselves and seek out in others.

But perhaps engaging with AI can be the middle path forward. As faculty, we can approach these AI podcast tools not as replacements for our presence, but as supplements that deserve careful, intentional integration. Doing so means setting clear boundaries around when synthetic voices serve learning and when they don't. The challenge here is getting people to pay attention and start seeing generative AI as something more than just ChatGPT.

We should pursue creating "AI-free zones" in our teaching practice —designated spaces and activities where students know they're engaging with their actual instructor's voice and presence. My guess is these authentic experiences will become all the more valuable in a landscape increasingly populated by synthetic alternatives.

Above all, we need to have open conversations with our students about the role of these tools. Rather than letting AI podcast features silently slip into our educational spaces, we should engage students in discussing how these tools affect their learning experience and their sense of connection with their fellow classmates.

The key challenge with generative AI in education isn't the technology– it's trying to find a footing amidst the onslaught of AI deployments so we feel confident understanding how these emerging tools might transform education. Preserving moments of genuine presence that remind students they're learning with real people who care about them, not just algorithms optimized for content delivery, should be the goal each time a new AI tool finds its way into our classrooms.