Me, myself, and (A)I: a Q&A with my AI avatar

When people ask for my first piece of AI advice, my answer is mostly the same: experiment with the available tools. Try ChatGPT. Have a conversation with Pi. Make art with DALL·E or Midjourney. Put 15 minutes on your calendar each day to play around with these tools. Start exploring how you can personally apply this technology to think and make in a better way. 

Over the last year and change, my team and I have been following this advice. I co-wrote my book Impromptu with GPT-4, created Fireside Chatbots with ChatGPT, and featured AI on nearly every episode of the Possible podcast to enhance and deepen the conversations. 

My latest experiment is creating an AI-generated version of myself. It’s not because I want another me—I actually thought I wouldn’t like doing this. But I was curious about testing the capabilities of a digital twin: how interacting with it might help me think differently, express myself in new ways, or connect ideas that I might not have otherwise. Since AI avatars will most certainly make a lot of people uncomfortable, I wanted to test what is possible and positive from this application of the technology. I think these avatars, when built thoughtfully, have the potential to act as mirrors to ourselves—ones that reflect back our own ideas and personality to examine and consider. 

So, a bit on how this was built: without getting too in the weeds, the video avatar is generated by Hour One. The voice cloning is done by 11ElevenLabs.  And the persona—as in, the way Reid AI formulates responses—is generated from a custom chatbot built on GPT-4 that draws from my books, speeches, articles, and podcasts that I’ve produced over the last 20 years (so even some of the ideas my AI avatar shares may be referencing older statements and thinking).

This experiment is still a work in progress, and I’m excited to continue to explore how AI and video can make content more dynamic and accessible. So the following is an interview with my AI avatar, in which we both take turns asking the other questions: https://youtu.be/rgD2gmwCS10

I’m interested in other positive, human-amplifying ways to use AI avatars. My favorite use case right now is translating my words from speeches, articles, and books into dozens of languages I do not speak, so that I can connect with people with whom I wouldn’t otherwise be able to communicate.

I’ve found the whole experience developing and experimenting with my AI avatar interesting and thought-provoking. As I’ve maintained from the time GPT first came out, I do not want nor expect this technology to replace humans. I view AI as a powerful amplifier of humans, and when I think about AI avatars through this lens, I begin to ask: What are some ways that I can use this tool to be more considerate and connected? More communicative, collaborative, and expressive? I think that we can and should use this technology to be and become more human.

A powerful technology has arrived. It’s important to be aware and thoughtful as we go into—and work to shape—the future.

Thanks for watching the video clip! I’d love to read what you think about it in the comments below.