Toying with AI: Adorable Chibi figures trend captivates creators worldwide

ChatGPT 4o‘s latest trend transforms users into adorable Chibi-style figures, capitalizing on the popularity of anime collectibles without the $10-100 price tag of physical figures. This phenomenon highlights AI’s evolving capability to create personalized digital memorabilia that closely mimics commercial products, while also raising questions about the environmental and privacy implications of casual AI image generation.

How it works: Users can create Chibi versions of themselves or others using a simple prompt in ChatGPT Plus 4o with just a clear photo.

• The prompt generates an image of a realistic glass gashapon capsule (similar to toy vending machine capsules) held between fingers, containing a miniaturized Chibi figure of the person.
• Unlike other AI image generation trends, this prompt focuses primarily on transforming the subject’s appearance and outfit without requiring extensive contextual information.

The technical process: The prompt instructs ChatGPT to create a specific style of collectible figure with recognizable characteristics.

• The Chibi style features exaggerated elements like compact bodies, oversized heads, and large eyes—making these digital figures instantly recognizable as part of the Japanese “small” or “short” aesthetic.
• The final image places the figure inside a plastic capsule reminiscent of Bandai’s Gashapon vending machine products, though such containers have been around for over five decades.

Real-world examples: The article’s author tested the prompt with several different subjects.

• The system successfully created Chibi versions of the author himself, Brad Pitt, and Mr. Rogers, focusing primarily on facial features and clothing details.
• The final images contain no recognizable backgrounds or accessories, putting all emphasis on the character representation.

Privacy considerations: The convenience of AI figure creation comes with potential data concerns.

• Users must upload their photos to OpenAI‘s servers, with no guarantees against those images being used to train future models.
• The process also consumes server-side electricity for both model building and prompt resolution, raising environmental considerations for frequent users.