Let’s be honest: the AI video space has exploded. Every week there’s a new tool promising Hollywood-quality results. But what happens when you combine AI image generation with video transformation and some creative editing? You get something that looks surprisingly real – and it’s easier than you might think.

I recently created a video that looks like I’m walking around, taking selfies with some of the most recognizable faces in tech. Sam Altman, Bill Gates, Steve Jobs – the whole crew. And no, I didn’t actually meet them at a conference. Here’s exactly how I did it.

The Concept

The idea was simple: create a video that simulates a casual “selfie tour” – you know, the kind where you’re at an event, bumping into people, snapping quick photos.

That iPhone aesthetic is key. Perfect stabilization but with that subtle smartphone look makes it feel authentic rather than overly produced.

Step 1: Generate the Images with Nanobanana

First things first – I needed the base images. I used Nanobanana to generate five different images, each featuring me with different tech personalities.

Why Nanobanana? It handles portrait-style images really well and gives you control over the composition. For selfie-style shots, you want:

• Natural lighting that looks like indoor/outdoor event spaces
• Slightly off-center framing (like a real selfie)
• Casual expressions and poses

I generated five distinct scenes, each with a different stars.

I found good prompts on this page: https://pixpretty.tenorshare.ai/de/ai-generator/selfie-mit-celebrity-ki-prompts.html

You can also use the new GPT Image 1.5 from OpenAI.

Step 2: Transform Images to Video with Artlist

Here’s where the magic happens. I used Artlist to create 5-second video transformations between each image. The key is the transition – going from Image A to Image B with smooth camera movement.

Each clip:

• Duration: 5 seconds
• Transition style: Smooth camera pan/walk movement
• Look: Stabilized but organic (that iPhone handheld quality)

The trick is making each transition feel like continuous movement. You’re not just fading between images – you’re creating the illusion of walking from one location to the next.

For this I use this prompt:

Do a smooth camera movement. First I take a selfie with Sebastian, then I walk further with a very stabilized image—but you can tell it's shot on an iPhone—to the next location where Sam is, and in the background you can see all the AI being developed by smart developers.

Step 3: Cut Everything Together in Canva

With all my clips ready, I moved to Canva for the final assembly. Say what you want about Canva – for quick video editing, it just works.

The editing process:

1. Select Instagram Reel
2. Import all 5-second clips
3. Add background audio
4. Add any finishing touches (slight color grading)

No need for Premiere Pro or DaVinci Resolve for this kind of project. Canva’s timeline editor handles the basics perfectly.

The Result

What you get is a video that looks remarkably like someone walking through an event, snapping selfies with different people. The AI-generated images provide the foundation, the video transformations create the motion, and the final edit ties everything together into a cohesive “story.”

Tools Used

Try It Yourself

The whole process took about an hour from concept to final video. The key learnings:

1. Be specific with your image prompts – The more detail you give about lighting, angle, and composition, the better your base images
2. Keep transitions short – 5 seconds per transition is the sweet spot. Longer feels draggy, shorter feels rushed
3. Simple editing wins – You don’t need complex software for this workflow

Is this the future of content creation? Maybe. Is it a fun way to spend an afternoon experimenting with AI tools? Definitely.

What creative video ideas are you experimenting with? Let me know in the comments!