Quick Dive: Making of Tales from the AI Frontier
Welcome to Deacs.AI Quick Dives! In these shorter segments, we zoom in on specific projects and use cases happening right here at Wake Forest. Today, host Ryan Scholl flips the script and interviews Michael Ferrari about the creation of the “Tales from the AI Frontier” video series – short, engaging explainers designed to make AI concepts accessible.
The Inspiration: Beyond the Talking Head
Michael explained the initial motivation was to move beyond typical screen recordings or standard “talking head” videos. He wanted something visually distinct and engaging, but also wanted to alleviate the pressure often associated with traditional video production.
“I was originally going to record just like videos of me… usual stuff that you see sitting at a computer, very YouTube style,” Michael shared. “But I thought that… it’s definitely for me a little bit more pressure to sit there and, you know, worry if during this take I suddenly scratch my nose… or fumbling over my words.”
The solution? Animation, powered in part by AI itself.
The Creative Process: A Blend of AI and Human Touch
Creating the animated videos involved a multi-step process, cleverly combining different tools:
- Character Creation (AI Assist): Michael started with an AI image generator. “You could kind of take your pick, whatever preference you have for like any kind of image generator,” he noted. He likely uploaded a reference image of himself and used prompts to generate the base cartoon avatar.
- Personalization (Human Touch): The generated image wasn’t the final product. Michael took the AI-created avatar and manually added personal details using traditional graphic editing software. “I went in and then I… added certain features that are a bit more me, so like maybe the facial hair, I put some of my like tattoos on there… putting the Wake Forest logo on my shirt.” This blend ensured the character felt unique and recognizable.
- Animation (Manual + AI Help): The animation itself was done in Adobe After Effects. Michael focused on simple movements – syncing the mouth to the audio and adding basic hand gestures – to keep it engaging without being overly complex. Crucially, when he hit technical roadblocks or needed to refresh his memory on After Effects techniques (like syncing animation to audio levels), he turned to AI like ChatGPT for help. “How would I time the movements of the hands and the… mouth moving with the audio? And then it gave me the little snippet of code that goes in there, reminded me where I put it, and in like two minutes, I had what I wanted.”
- Voiceover (Human): The narration is Michael’s own voice. He mentioned enjoying voiceover work and found it less pressure-filled than being on camera. (Though the potential of AI voice tools like 11 Labs was discussed as a future possibility).
- Scripting (Iterative): Michael’s scripting process was organic. He’d hit record, talk through the topic freely (“really sharing my story and my opinion”), and then refine that recording into a tighter, more structured script for the final voiceover.
Choosing the Topics: Practical Applications
The video topics weren’t random; they stemmed directly from Michael’s own primary uses for AI:
- Prototyping: “One of my favorite ways to use it… I can’t even count the number of projects that I’ve been able to start” using AI to quickly test concepts.
- Learning & Refreshers: A common and powerful use case, drawing on his experience using AI to understand new things or recall forgotten skills (like the After Effects techniques).
- Graphic Design: Another practical area where AI can assist, even if results vary across different image generation tools.
Why This Format?
The goal was to create something easily digestible (“quick dives”) that could complement longer podcast discussions or stand alone. The animated format offered a unique visual style while the AI assistance streamlined parts of the production process, making it a prime example of using the very technology the series explores.
Key Takeaway:
The “Tales from the AI Frontier” project showcases how AI tools can be creatively integrated into content creation. It’s not about fully automating the process, but about leveraging AI for specific tasks – like initial image generation or technical troubleshooting – while still applying human creativity, personalization, and judgment. It demonstrates that with a bit of experimentation and a willingness to learn (sometimes from the AI itself!), engaging multimedia content is becoming increasingly accessible.