Legal Aspects to be Considered When Joining Viral Trends

Ana Cristina Ortega, Associate at ARIAS Guatemala and an expert in Intellectual Property and Technology Law, presents this article on the importance of evaluating the risks associated with viral trends on social media. 

Recently, a social media trend went viral involving the transformation of personal photographs into caricature-style images "in the Ghibli style." Thousands of individuals across the globe uploaded their photos to GPT-4º, which then converted them into caricatures that mimicked the aesthetic of Studio Ghibli animations. Shortly thereafter, a significant debate emerged centered around two key issues: first, the alleged unauthorized use of Ghibli images to train the AI model and generate the resulting outputs; and second, the potential risks associated with sharing personal photographs containing biometric data with a public AI tool, without a deeper understanding of the implications of voluntarily "surrendering” such information to an artificial intelligence system. 

This article explores both topics, with the objective of enabling readers to make informed decisions about whether or not to participate in viral trends, even at a superficial level, and whether such participation could entail potential risks. 

First, we must examine the alleged unauthorized use of Ghibli images to train or feed an artificial intelligence system. Copyright law protects the rights of creators, artists, and authors over their literary and artistic works. While copyright statutes typically do not provide exhaustive lists of protectable works, it is essential to highlight that copyright protects the expression of ideas, but not the ideas themselves. Thus, styles—whether musical or artistic—are generally not protected, to avoid monopolization of entire styles that could stifle the artistic expression of others. However, it is crucial to distinguish between what constitutes a "style" and what qualifies as a "work" that possesses specific, distinctive, and identifiable elements capable of protection. 

In light of the above, it could be concluded the images and animations created by Hayao Miyazaki, founder of Studio Ghibli (a Japanese animation studio, considered as one of the best that has produced some of the greatest box office movies in Japan´s history as "Spirted Away or Sen to Chihiro no Kamikakushi in Japanese), are protected under copyright law, as they constitute the author’s original and characteristic expression, attributable to the creator, so it is up to its owner to weigh up the legal actions that may be taken. The viral images were produced using a new ChatGPT update that enables users to transform images and personal photographs to appear as though they were created by Miyazaki or as part of Studio Ghibli’s animated films. This sparked an ongoing debate and analysis for the purpose of this article, centered first on ethical considerations inherent in discussions on AI—namely, the use (sometimes unauthorized) of copyrighted works to train AI models. This issue has already led to several lawsuits filed by artists alleging that AI-generated works closely resemble their own, as the original copyrighted content was allegedly used without permission during the AI training process. The second issue involves potential infringements of intellectual property rights, particularly copyright, raising concerns that users may unknowingly violate third-party rights when generating content with these tools—especially if the material used to train the model is improperly licensed or entirely unauthorized. 

A third and equally pressing concern relates to the voluntary submission of biometric data to AI tools, without a full understanding of how such information may be stored, used, or repurposed. Data protection laws in jurisdictions with effective legal frameworks require that entities collecting, storing, and processing personal data comply with various legal obligations, including ensuring that personal data is lawfully obtained and used solely for specific, previously defined purposes. 

When a user uploads a personal image to an AI system for purposes such as cartoonification, several risks arise. Most notably, the user relinquishes control over the future use of their image. These images may be used to further train AI systems or even be exploited to identify individuals for purposes such as fraud, scams, identity theft, or unauthorized commercial use, including the sale of data to third parties. 

Although data protection laws often require user consent and a clearly defined, legitimate purpose for data collection and processing, AI tools frequently draft vague and ambiguous terms and conditions. Users—who often do not read these terms in detail—may unknowingly agree to terms allowing their photographs (and thus their biometric data) to be used for broader or undisclosed purposes, including model training. 

Some commentators have compared this to the act of uploading personal photographs to social media, questioning whether the risks are equivalent. While it is true that uploading photos to social platforms exposes certain personal data and implies a level of consent to the platform's usage policies, these networks tend to be more heavily regulated. Their terms and conditions are generally more transparent regarding the scope and purpose of data collection and processing, and they typically offer privacy controls that allow users to restrict how their content is used. As such, while there is still a degree of risk involved in sharing content on social media, it is relatively lower than the risks posed by submitting personal data for AI training purposes. 

In today’s digital world, it is nearly impossible to avoid using AI tools or engaging with viral trends. What is essential, however, is the ability to identify and assess the potential risks associated with uninformed use. In this regard, we recommend the following: 

While many of these viral trends are entertaining and offer increased visibility, it is crucial to carefully assess the hidden risks involved—both for users who unknowingly relinquish personal data and for creators whose intellectual property rights may be infringed. Such risks, though often silent or overlooked, may result in serious legal consequences and significant violations of proprietary rights. 

The information provided by ARIAS® is presented for informational purposes only. This information is not legal advice and is not intended to create, and does not constitute, an attorney-client relationship. Readers should not act upon this information without seeking advice from professional advisers.