TLDR

• OpenAI’s GPT-4o includes a powerful new AI image generator that’s gone viral for creating Studio Ghibli-style images
• The tool can replicate studio styles like Studio Ghibli but raises copyright concerns as courts debate whether AI training on copyrighted works is legal
• Several lawsuits against AI companies like OpenAI claim copyright infringement for training on works without permission
• The new image generator excels at text rendering, precise prompt following, and maintaining consistency across multiple images
• OpenAI’s tool produces more accurate style replicas than competitors like Google’s Gemini and xAI’s Grok

OpenAI recently launched a new image generator as part of its GPT-4o model, and it has quickly gone viral on social media. Users have flooded their feeds with AI-generated images mimicking the distinctive style of Studio Ghibli, the famous Japanese animation studio behind beloved films like “My Neighbor Totoro” and “Spirited Away.”

In just 24 hours since the feature went live, people have created images showing Studio Ghibli versions of Elon Musk, “The Lord of the Rings,” and President Donald Trump. Even OpenAI CEO Sam Altman appears to have changed his profile picture to a Ghibli-style image likely created with the new tool.

The feature allows users to simply upload existing images into ChatGPT and ask the system to recreate them in various styles. This ease of use has contributed to its rapid spread across social media platforms.

This release follows Google’s similar feature in its Gemini Flash model, which sparked its own viral moment earlier in March when users discovered they could remove watermarks from images. Both tools make it easier than ever to recreate the styles of copyrighted works with simple text prompts.

These new AI image tools have reignited concerns about copyright infringement. Several lawsuits against AI companies claim these systems were trained on copyrighted works without proper permission or compensation.

AI Art & Legal Gray Areas

Evan Brown, an intellectual property lawyer at Neal & McDevitt, explains that these AI image generators operate in a legal gray area. While style itself isn’t explicitly protected by copyright law, the way these systems learn those styles might be problematic.

“I think this raises the same question that we’ve been asking ourselves for a couple years now,” Brown said in an interview. “What are the copyright infringement implications of going out, crawling the web, and copying into these databases?”

The New York Times and several publishers have active lawsuits against OpenAI. They claim the company trained its AI models on copyrighted works without proper attribution or payment. Similar claims have been made against other AI companies, including Meta and Midjourney.

OpenAI stated that while ChatGPT refuses to replicate “the style of individual living artists,” it does permit recreating “broader studio styles.” However, this distinction becomes blurry when considering that living artists like Hayao Miyazaki, co-founder of Studio Ghibli, pioneered their studios’ unique styles.

studio ghibli is out, dr seuss is in pic.twitter.com/4ECxwLLkoj

— Jordi Hays (@jordihays) March 26, 2025

Users have also been able to recreate styles from other well-known sources. Examples include portraits in the style of Dr. Seuss and wedding photos reimagined in Pixar’s distinctive look.

Tech journalists compared several popular AI image generators, including Google’s Gemini, xAI’s Grok, and Playground.ai. They found that OpenAI’s new tool created the most accurate replicas of Studio Ghibli’s animation style.

According to OpenAI, the new image generator was designed to be both “beautiful and useful.” The company claims their tool excels at accurately rendering text, following prompts precisely, and leveraging GPT-4o’s knowledge base.

A key advantage is the generator’s ability to maintain consistency across multiple images in a conversation. This makes it easier to refine images through natural dialog, ensuring that elements like characters remain coherent as users make adjustments.

The model also demonstrates strong “instruction following,” handling detailed prompts with up to 10-20 different objects while maintaining their traits and relationships. This allows for better control than previous systems, which typically struggled with more than 5-8 objects.

OpenAI has implemented safety measures in the new tool. All generated images include C2PA metadata identifying them as AI-created. The company also blocks requests that may violate their content policies, with heightened restrictions when real people are involved.

The new image generator is now available to Plus, Pro, Team, and Free users as the default image generator in ChatGPT. Access for Enterprise and Education users is coming soon. Developers will be able to generate images with GPT-4o through an API in the coming weeks.

OpenAI acknowledged some limitations with the current system, including occasional image cropping issues, inaccuracies with multilingual text rendering, and challenges with precise editing. The company says it is working to address these limitations in future updates.

The success of this new image generator seems to be driving high demand. OpenAI delayed the rollout to free-tier users on Wednesday, citing the overwhelming number of people trying to use the feature.

While the technology represents a major advance in what AI can create, we’ll have to wait for court decisions to determine if these systems violate copyright laws.

The post GPT-4o Image Feature Creates Studio Ghibli-Style Art Amid Legal Uncertainty appeared first on Blockonomi.