Intro 

We’ve all seen headlines about AI taking over creative work. But what does it really look like to hand over the entire creation of a learning video—script, visuals, narration, even translation—to AI? 

In this piece, I sat down with my colleague Zach Anderson, an instructional designer and videographer on our team, to unpack what it really took to build an AI-powered cybersecurity training video for a medical supplier. The challenge? Create a training video, entirely with AI tools, that could be translated into five languages and still meet the company’s quality, brand, and technical standards. 

It was a real client, a real deadline, and a real test of what AI can (and can’t) do in creating learning media. What Zach discovered surprised him. And it reinforces a truth we keep coming back to: tools don’t create impact, people do. 

This was not a theoretical project. 

Robin: So, give me the short version of how this project got rolling. 

Zach: It started out very differently than how it ended. Initially, we were going to make a longform cybersecurity awareness video, and I had this wild idea of doing a noir theme. You know, something like “you wouldn’t steal a password,” with a gritty voiceover.  While translation and the international audience were important considerations, the real driver of the shift was the client’s request that we use some specific AI tools for production. That request—not just creative preference or logistical constraints—completely changed the design direction.” 

Robin: And that’s when AI became essential? 

Zach: Exactly. The design constraints and the turnaround time made traditional video production unrealistic. We pivoted to using various AI assisted tools, including HeyGen for talking head avatars, Adobe Firefly for visuals, ChatGPT for prompt generation and script iteration, and After Effects to stitch it all together. I was working in tandem with our client contact, who was also using a similar set of AI-powered tools. We needed everything to match visually across both projects. 

The tools were powerful, but the real work was human. 

Robin: What was it like to produce an entire video using AI? 

Zach: Honestly? It was emotional. I have strong feelings about AI. Specifically, about the ethics, the implications to personal creativity, and I wasn’t sure how I’d feel actually doing the work. But I saw it as a chance to level-set what I thought I knew about these tools versus what they could really do in a real-world project. And it totally reshaped my assumptions. 

Robin: In what way? 

Zach: Well, some tools were incredible. ChatGPT was a game changer for writing and refining prompts that helped me to more accurately get the tools to do what I wanted. But even though all the assets were AI-generated, the actual creative labor still felt very human. I was constantly iterating, troubleshooting, trying to keep characters visually consistent, and solving translation issues. It felt like a video shoot day, honestly, just without a camera in my hand.