I’m laughing my butt off. I asked ChatGPT to make me into an action figure so I could ride the wave of the viral AI trend, and wow, did it deliver. It turned me into a rock star figurine in skin-tight leather, suitable for ages 4 and up.
The amusing action figure trend has expanded across social media. It’s also called the “Barbie box” trend.
It’s tough to pinpoint the origin, but you’re probably seeing your friends, family and colleagues sharing images of themselves reimagined as packaged toys.
Here’s how to use an AI chatbot to make an image of your own, and how to 3D print it so you can actually own that personalized action figure in real life.
How to make an AI action figure of yourself
Step 1: Pop on over to ChatGPT, which now offers free image generation for everyone. I asked ChatGPT how many images it would let me make free, and it said it would do up to 15 per day. After that, image generation goes on pause for free users until the limit resets.
Step 2: I queried ChatGPT for guidance on how to make me into an action figure, and it asked for a full-body photo. You can upload images from the “ask anything” box. I found a not-very-clear shot from my photo archives and sent it along. Good enough, I figured.
Enlarge Image
This is the full-body image I gave ChatGPT to work with.
Step 3: The AI service asked me for preferences for outfit or costume, pose or expression, accessories or weapons, name or title for the action figure and packaging style — such as retro, modern or comic-book inspired. You can be as detailed or vague as you like. I went with vague: “I’d like my outfit to look like a rock star with a rocking pose, a guitar and a retro packaging style.”
Step 4: Wait, then share. It can take ChatGPT a few moments to generate your action figure. If you like what you see, then go ahead and download it and blast it across your socials.
Enlarge Image
Here’s the full glory of my AI action figure.
Step 5 (optional): If ChatGPT doesn’t quite nail it, then ask for tweaks. I tried a new version with blue jeans and a pink Telecaster guitar. It delivered, but forgot the action figure part and the new image looked further removed from my actual facial appearance. A third tweak took me into Billy Idol territory when I asked for spiky blonde hair. My action figure’s leg ended up outside of the “plastic” packaging window. I tried again.
Enlarge Image
This looks nothing like me.
Don’t expect perfection. The more I asked for changes, the weirder it got. Eventually, ChatGPT turned me into a Jon Bon Jovi-esque figure with three fingers on my left hand playing a guitar with seven strings.
How to 3D print your creation
Our resident CNET 3D printing expert James Bricknell has taken this craze one step further and, using some clever AI technology, 3D-printed one of these ChatGPT images. He’s here to explain the process.
As soon as I saw the first action figure image from ChatGPT, I knew I wanted to create a 3D-printed version. Turning an image into a 3D model can be complicated, sometimes involving large camera rigs and a lot of clever software, but the latest AI advancements mean creating something simple and fun is a lot easier. When CNET editor at large Scott Stein posted his action figure on Threads, I figured it was a good place to start.
Step 1: Editing the image. Your first image is likely to be a figure in a blister pack. That looks amazing, but a lot of the excess artwork won’t translate when 3D printing. You’ll need to crop out as much extraneous material as you can, focusing on the model with a solid color background if possible.
You may find you need to tweak the image even more to get a good 3D render. I had to draw a line between Scott’s glasses and eyebrows so the modeling software could recognize them as two separate objects.
Step 2: Access Maker World’s Maker Lab. Bambu Lab makes some of the best 3D printers available today, and it has an excellent website for finding 3D models to print. Maker World has a tab called Maker Lab, where the company has a host of interesting AI projects.
The one we want is down at the bottom, called Image to 3D Model. You will need to create a login to use this feature, but it’s free to sign up.
Step 3: Import your image. Once you click the Try it button, you’ll be asked to upload the cropped image from step one. When you do, the AI will kick in to create something 3D printable. This takes a few minutes and may not give you the perfect image on your first try. If you don’t like the output, you can try a different image or remove the background from your image to make the model pop.
This process costs Maker World credits, but you start with some as part of your sign-up. Thankfully, you only pay the credit when you download the finished model. This means you’re free to try multiple different renders to get it right. When you’re happy, click export to download the model.
Step 4: Importing to your slicer. The downloaded file has a .OBJ file format, which is recognized by pretty much every 3D printing slicer out there. I used Bambu Lab’s Bambu Studio, but you could use Prusa Slicer or any other slicer for your specific 3D printer. Most of the slicers out there have a repair function and it may be worth running your model through it. I didn’t encounter any errors on my tests, but better to be sure.
If you have a printer with a color system now would be the time to paint you desired colors onto your model. If you don’t, then it will just print as a single color object.
Note: Your model may not have flat feet. I noticed this model didn’t, so I had to make sure the supports were active on the soles of his feet to keep him upright.
Step 5: Send your model to the 3D printer. Once you’ve done all the other steps, it’s time to print. I was using the new Bambu Lab H2D and it took a little over 11 hours to get a full-color 3D print. A single-color print would take significantly less time. I also increased the size of the model by 150% as the Maker World output was fairly small.
You can see I used a lot of support to make sure his arms and glasses didn’t break off in the printing process. At this point, you can start thinking about potential accessories. I printed a small VR headset to go with Scott’s figure that I glued to the hand. I think it rounded it off well.
Step 6: Bask in your new Mini Me’s presence. That’s it. You now have a small action figure based on your ChatGPT image. It didn’t take long, and using the Maker Lab AI saved you potentially hundreds of hours learning how to model it yourself. It’s not perfect, though. A professional 3D modeler could do a much better job, and make it articulated to boot. But for a fun little project, this is excellent.
Inspiration for your AI action figure
Need some inspiration? Check out these fun and funny AI action figures shared on social media.
The UK’s Royal Mail took to Instagram to share its adorable “Postie” action figure, complete with an envelope and an optional safety vest.
Enlarge Image
The Royal Mail made an AI-generated Postie action figure.
Seinfeld fan account Seinfeldism generated a series of Seinfeld-related images. The Kramer is pitch-perfect.
Enlarge Image
This AI-generated Kramer action figure looks like it should be official merch.
Barbie boxes aren’t just for humans. Bear, a German shepherd with 134,000 Instagram followers, looks pretty sweet as a toy.
Enlarge Image
Dogs make good action figures, too.
ChatGPT is reasonably skillful at taking a full-body photo and transforming it into a fake toy. If nothing else, it should give you a good laugh. Get creative with it. I play in a rock band. ChatGPT was surprisingly accurate: My action figure was dressed the way I dress on a gig night, even though it made me look a little like a female Bruce Springsteen. Rock on, ChatGPT.
