You may have noticed all of your friends on social media are turning themselves into Barbie dolls and action figures.
It’s a viral trend of 2025 that is powered by artificial intelligence technology now at the general populations finger tips via popular apps and online tools.
The latest wave in AI art generation comes from the enduring popularity of “Barbiecore” aesthetic, a hyper-feminine aesthetic inspired by the iconic Barbie doll, revitalized in 2023 by the “Barbie“ movie featuring Margot Robbie. “Barbiecore” is a style characterized by bright pinks, bold accessories and looks that would notably be worn by Barbie.
Social media users are using tools like ChatGPT and image generators available through Canva to turn their own personal photos into action figures or Barbie dolls, complete with accessories of their choice that come encased in original plastic packaging. No single person is credited with starting the viral AI trend, but it’s clearly a standout social media firestorm of 2025, using hashtags #BarbieBoxChallenge and #AiBarbie .
WSLS 10 News anchors are even having fun with the trend.
Of course we had to try it. We got some pretty fun results with the evening team. We used variations of this prompt:
“Create a realistic action figure toy of the four people in this photo, the WSLS evening news anchors. Pictured right to left is meteorologist Jeff Haniewich, news anchor John Carlin, news anchor Rachel Lucas and Sports Director Eric Johnson. The figures should be a full figure and displayed as a set in a blister pack. They always have a news gathering camera, a reporter notepad, an anchor desk and a coffee mug and a microphone.”
How to create your own action figure or Barbie
Start with an AI prompt:
Open Canva AI Image Generator or ChatGPT (available in ChatGPT Plus and Pro plans only) with image generation enabled and write a prompt similar to this:
“Create a realistic action figure toy of the person in this photo, (NAME). The figures should be a full figure and displayed in a blister pack. They always have these accessories (enter accessories here).”
*Note: Using the word “Barbie” in your prompt may violate AI’s policy. That’s why we used “action figure” or “doll.”
Upload your photo:
Upload a reference photo or selfie to better guide how the image looks. A high-resolution image is recommended. Use the chat to tell the AI what elements of the photo should be kept, such as hair color, outfit or pose.
Review and Regenerate if needed:
ChatGPT or Canva will create an image based on the prompt. If the result isn’t exactly what you are looking for, tweak the description directly in the chat—for example, by adjusting the background, length of hair, adding accessories or changing the packaging design.
Download and post to social media:
Download the image to your device and post it on social media. Don’t forget to include the #AIBarbie and #BarbieBoxChallenge. This trend is particularly popular on Facebook, Instagram, TikTok and Threads.
What you’re not seeing on social media
Just like real life toys, there are some rejects that come off the assembly line. You’ll likely find a lot of those through trial and error on your Barbie-making journey, too. You might not being seeing those on social media. We noticed AI still has difficulty with wording on images and created several misspellings. This is often easier to correct with image editing tools.
AI also had a little difficulty creating multiple dolls from a group photo we submitted. In several generations it mixed up our clothing, hair colors of even genders. Sometimes it left out a hand or other body part. Either way, the reject photos gave us a good laugh.
We found AI did a much better job of generating one doll at a time.
Lets have some fun! We would love to see your AI dolls -- and reject images!
Submit them to WSLS through Pin It.