What it is
The mass craze for pushing any photo — selfies, memes, film stills, news shots — into the recognisable animation aesthetic of Studio Ghibli (Miyazaki). The phenomenon is not a "filter" but proof that recent models can hold a coherent artistic style over an arbitrary image from a single sentence.
Where it came from
On 25 March 2025 OpenAI built native image generation straight into GPT-4o. Within days X and Instagram feeds were flooded with "Ghiblification". The head of OpenAI publicly joked that "our GPUs are melting" under the load, and temporarily introduced limits.
Why it took off
- Emotion plus nostalgia. A warm, recognisable style turns an everyday photo into "a frame from a feel-good cartoon" — that is something people want to repost.
- Zero barrier. No Midjourney, no prompt engineering: "make this photo in Ghibli style", right in the chat.
- Network effect. One template → an avalanche of copies and parodies → media coverage → another round.
How to use it today
- Any strong image generator (GPT Image, Nano Banana, Midjourney, FLUX.2) plus a line like:
redraw this photo in 90s studio 2D animation style: soft watercolour shadows, warm light, kind faces. - Don't name living artists or studios for commercial work — describe the traits of the style (palette, texture, lighting) rather than the brand. What you get: a series of recognisable illustrations for avatars, covers, cards and stories.
What to watch out for
- Ethics and law. The trend reopened the argument: imitating the signature style of living artists is widely seen as unethical; Miyazaki has spoken out sharply about AI. For client or commercial work this is elevated legal risk.
- In production, "the style of a specific studio" is better replaced with a neutral description of the aesthetic.