Bored of Ghibli? Try 5 unique AI portrait styles as ChatGPT’s new image generator goes free

The trend for Studio Ghibli-style AI portraits shows no signs of slowing down anytime soon, with even popular world leaders, actors and cricketers joining the chorus. However, the native image creation feature that powers the Ghibli-style AI images, which has only recently been made available to free users, is capable of creating much more than just the Japanese anime style. Here’s a look at the 5 unique styles you can give your portraits.

1) Lego style:

Lego-style images add a playful, blocky character to your portrait shots, making them look like the design of iconic toys, that also have multiple movies named after them.

2) Pixar:

Pixar-style images feature large and expressive eyes, soft lighting, smooth texture, and detailed shading, making the portrait shots look like they are straight out of an animated film from the popular studio.

3) Southpark:

South Park style effect will transform the people in the portrait shot into a flat, cutout like character with simple shapes, thick outline and facial expressions. The ultimate output will be an image with cartoonish and satirical look.

4) Comicbook:

Comicbook style adds bold lines, dramatic shading and high contrast colours to the portrait shots, making them look like they were taken straight out of a popular comicbook.

5) Simpsons:

Simpon style effect will turn your images into yellow-toned skin with overbite smiles and exaggerated facial expressions. The ultimate output will make the persons in the image look like they are residents of Springfield – the town where The Simpsons is set.

If you’re still not convinced about switching to one of the above styles, there’s always the option of following the crowd and going with Ghibli-style images.

How to prompt ChatGPT for best images?

With ChatGPT’s new image generator, which can refine images through natural conversation, you can’t get much wrong when writing prompts for the chatbot. But there are a few basic tips to keep in mind when writing such prompts to get the best possible output.

Give ChatGPT details about style elements for the image, such as colours or expressions, and describe the background, lighting, textures and emotions for the image.

Here’s an example of a detailed prompt generated by ChatGPT itself, “Generate a Pixar-style 3D animated portrait of a young woman with large, expressive eyes, soft lighting, and smooth textures. The image should have a warm, cinematic color palette and a slight glow effect.”

What is native image generation? Why is it leading to great pictures?

ChatGPT has been able to convert text to images for some time, but this has been done using external models such as OpenAI’s DALL-E 3. However, OpenAI recently unlocked the native image generation capabilities in GPT-4o, the large language model that runs ChatGPT.

This has led to the ability to generate and edit images directly, taking advantage of its multimodal capabilities. The resulting images are better at handling objects and are more accurate, while also being able to take advantage of ChatGPT’s text knowledge.