OpenAI’s latest update to ChatGPT’s image-generation capabilities has ignited both creative fervour and ethical scrutiny. The platform’s GPT-4o model, now accessible to paid subscribers, can produce Studio Ghibli-inspired visuals with unprecedented fidelity, flooding social media with dreamlike landscapes and characters reminiscent of Hayao Miyazaki’s films.
However, this technological leap coincides with a controversial relaxation of content moderation policies, allowing AI-generated images of public figures, racial features, and hateful symbols under specific conditions.
OpenAI’s model behaviour lead, Joanne Jang, framed the shift as a move away from blanket restrictions toward a “more precise approach focused on preventing real-world harm.” The company now permits users to opt out of being depicted in AI-generated images, rather than unilaterally blocking requests involving public figures like Donald Trump or Elon Musk.
Under the revised guidelines, users can generate hateful symbols in educational or neutral contexts, provided they do not “clearly praise or endorse extremist agendas.” Requests to modify physical traits, such as altering eye shape or body weight, are also now permissible. However, safeguards remain stricter for images involving children compared to OpenAI’s previous DALL-E 3 tool.
While ChatGPT can emulate studio aesthetics like Pixar or Studio Ghibli, it prohibits replicating individual living artists’ styles—a stance that has reignited debates about AI training datasets and fair use. OpenAI’s free-tier users, restricted to DALL-E 3, encounter stricter policies, with the platform explicitly blocking Ghibli-style requests due to copyright concerns. Paid subscribers, however, face fewer limitations, though OpenAI maintains a conservative approach to avoid infringing on artists’ work.
For users without access to GPT-4o, platforms like Gemini, GrokAI, and Craiyon offer Ghibli-inspired image generation with precise prompts. Advanced tools such as Runway ML and Leonardo AI provide finer control over details like colour palettes and character designs, though some features require paid upgrades.
OpenAI’s adjustments mirror moves by Meta and X to permit more controversial content to balance user freedom with responsible governance. Google faced backlash earlier for Gemini’s multiracial depictions of historical figures.