This case reminds us that it is essential to overcome the threshing of the media, especially when it can threaten our fundamental rights, in particular the protection of privacy.
During the last days, the images generated by the last OpenAI tool have social flood networks. These creations, strongly inspired by the style of Ghibli’s study, famous for films like my neighbor Totoro or Chihiro’s trip, triggered a viral phenomenon called “Ghibli effect.” If this trend is seduced by its aesthetics, it raises crucial questions about copyright and confidentiality of personal data.
When AI reproduces protected styles
From a legal point of view, the simple fact of reproducing an artistic style does not constitute, in itself, a violation of copyright. However, the ability of an AI to generate images in a specific style with disturbing fidelity raises an important question: Have you been trained in protected works?
Has Operai obtained licenses from Ghibli Studio or other creators to use their works as a learning base? Nothing allows you to affirm it. In the United States and other places, legal procedures are increasing against Operai and Anthrope for copyright violation. The authors, artists and creators denounce the indisputable use of their works by AI models, without compensation or prior agreement.
This debate goes far beyond the only artistic framework. If an AI can recreate images in a protected style, Where is the border between legitimate inspiration and abusive reproduction?
Personal images in free access: the implicit consent trap
The “Ghibli effect” not only raises a copyright problem: it also has Important implications for the protection of personal data.
In the European Union, OpenAI justifies the training of its models in images recovered online invoking the “legitimate interest” (article 6.1.F of the GDPR). However, this legal basis requires strict obligations, in particular the implementation of guarantees to avoid the abusive exploitation of confidential data.
But another problem arises: consent. To obtain their own Gibli -style portrait or another style, thousands of Internet users voluntarily download their photos on the platform. However, in doing so, they accept the general conditions of OpenAI and authorize the company to exploit its images (article 6.1 of the GDPR).
The result? Operai wins a large new faces base for free No need to extract them from social networks. Worse, these photos can include private photos, such as family images or intimate photos, which were initially not accessible on the Internet. Meanwhile, competitive platforms can only see the “fiberized” version of these images, which gives Openai a considerable strategic advantage.
Towards a stricter supervision of generative
The consequences of this situation are not trivial. On the one hand, artists find a system capable of imitating their work without their saying. On the other hand, users involuntarily provide the platform for a precious resource: their personal data.
Given these challenges, regulators around the world are beginning to react. HE European Data Protection Committee He recalled in his opinion 28/2024 that these platforms must imperatively respect the principles of the GDPR. Other jurisdictions, such as some US states, provide stricter regulations to supervise the use of personal data by AI.
But will these efforts be enough? Artificial intelligence evolves faster than legal frameworks and Abusive exploitation risks persist. As long as clear and effective safeguards are not implemented, the legal vacuum will be an opportunity exploited by companies to the detriment of artists and user privacy.
The “Ghibli effect” highlights the growing tensions between artificial intelligence, copyright and protection of personal data. On the one hand, artists see their reproduced style without control. On the other hand, Internet users involuntarily deliver sensitive information to an increasingly voracious.
These questions are not anecdotal: they draw the future of creation and privacy in the AI ​​era. Far from being a simple aesthetic tendency, This phenomenon illustrates the main challenges that legislators and regulators will have to assume to protect the rights of creators and citizens against new technological advances.
