While this trend has captivated many with its artistic appeal, it raises important legal concerns that remain largely unnoticed by the general public.

One significant issue is the potential for copyright infringement, as OpenAI’s AI models are trained on datasets that include publicly available data, which may encompass copyrighted works. Additionally, users may unknowingly be contributing their images to further train these models, raising questions about data privacy and consent.

When asked about its position on the growing trend of AI-generated images styled after Studio Ghibli's iconic animation, ChatGPT acknowledges that the practice "can be a gray area" in terms of copyright law. It further suggests that in order to avoid copyright infringements, users could frame their prompts with phrases such as "inspired by," avoiding direct replication of copyrighted characters or works.

But is this all there is to it? Would it be true that as long as the output does not imitate the original work, it would not be considered as a copyright infringement?

There are two main areas of contention in the AI art creative process where copyright could be infringed.

First, in the input stage, when the generative AI systems are trained using vast amounts of data coming from both publicly available data and copyrighted works. Second, in the output stage where the works generated by AI could or could not be copyrighted due to the difficulty of identifying the creator and owner of the produced work.