The article also details a user's experience with different AI models attempting to generate images in the style of Studio Ghibli, Jeff Koons, and Andy Warhol. While ChatGPT blocked requests for Ghibli and Warhol styles, it allowed a Koons-style image. Grok 3 successfully generated images in all three styles, albeit with varying accuracy, while Claude 3.5 Haiku struggled to replicate the desired styles, offering more abstract interpretations. The differing responses from these AI models highlight the ongoing challenges and inconsistencies in AI-generated art and content policy enforcement.
Key takeaways:
- OpenAI has blocked some users from generating images in the style of living artists, including Studio Ghibli, using its image generation model.
- The internet saw a surge of Studio Ghibli-style images after ChatGPT's image generator update, prompting discussions about copyright and AI-generated art.
- Different AI models, such as Grok and Claude, produced varied results when attempting to generate Studio Ghibli-style images, with some models adhering to content policies more strictly than others.
- Attempts to generate images in the style of other artists, like Jeff Koons and Andy Warhol, showed inconsistent enforcement of content policies across different AI platforms.