By The Editorial Board
Recently, OpenAI released a feature that allows users to generate images from text prompts. One of the most popular trends has been creating images in the style of Studio Ghibli, the Japanese animation studio is known for its spectacular hand-drawn animation. These AI images replicate this style of art, starting a phenomenon that is not only harmful to the environment for its massive energy usage but also undermines the creativity that made Studio Ghibli special in the first place.
On March 31, the ChatGPT–4o model became available to all users for free, bringing Studio Ghibli into the hands of billions worldwide. One of the reasons this tool became popular is that it was a lot more accurate and responsive to user prompts compared to other AI tools. This allowed users to generate images of adult figures and was less likely to reject user prompts for being offensive, creating more open-ended and sometimes offensive creations.
Even the White House joined in the trend as they posted a photo on X, showing an ICE agent arresting Virginia Basora-Gonzalez, an illegal migrant and fentanyl dealer, through the studio Ghibli feature.
Sam Altman, CEO of OpenAI, posted a Studio Ghibli photo of himself, joking that the high volume of image generation will melt the OpenAI servers. This skyrocketed the trend’s visibility and significantly boosted ChatGPT’s user engagement.
What are the pros of image generation from AI?
Art can be more widely accessible and applicable in a variety of contexts. It’s not a mystery why the Studio Ghibli trend is so popular. People love to see how their lives intertwine with the beautiful, crafted style of Hayao Miyazaki, the creator behind Studio Ghibli.
For creatives, AI can lead to more ideas, as you don’t need to rely solely on your imagination but on visible images prompted by text. AI changes the game with what it means to be an artist, but it begs the question, what are you replicating and not creating? Art has always been accessible — all you need is your imagination. This makes art easy, giving an end result without all the work, talent and dedication.
What are the downsides of image generation from AI?
First, much of the art that OpenAI is trained on is copyrighted material. Many of the famous hand–drawn works by Studio Ghibli, such as “Spirited Away” and “Howl’s Moving Castle,” are used as training for large language models to replicate, even though they do not own the rights to the works. It is not a good precedent to appropriate work from other creators, even if you think more art will be created through AI. Intellectual property is protected in every field and service; however, for some reason, AI large language models have been able to use intellectual property for their own benefit because it hasn’t been properly regulated yet. Many argue that ChatGPT isn’t “creating” anything, it is simply plagiarizing it and calling it their own. As Sam Altman defends his product by highlighting its “fair use” under the U.S. Copyright laws, we begin to understand the ethical and moral dilemma involving ChatGPT’s use of artists’ work without their consent.
However, OpenAI took a step back after receiving backlash for the Studio Ghibli image generations. On April 2, OpenAI started restricting Studio Ghibli requests and limited the number of image generations per day for users using the free version. The company stays diligent to not end up with a large copyright lawsuit for its “fair use,” as it rides the line of what is legal.
Secondly, OpenAI is harmful for the environment. Reportedly, in a conversation of 20-50 questions with ChatGPT, 500 ml of water is consumed during the conversation, making the potential water footprint daunting as the number of users continues to rise. The main contributing factors are the sheer amount of energy required to train large language models and the amount of water needed to cool their systems. Image generation requires significantly more processing power than just a text prompt, highlighting that we need to draw the line of when image generation is actually useful. As users should remain mindful of when using such features are appropriate, it is also up to OpenAI to be transparent about its ecological footprint. AI will continue to be more and more prevalent in our daily lives; so, we need to enforce that these companies stay accountable for the environment that we share.
