top of page

By:

Correspondent

21 August 2024 at 3:50:16 pm

Algorithmic Alibi

Meta’s handling of child sexual-abuse material has exposed the hollowness of the tech behemoth’s claim to be merely a simple intermediary. A company whose algorithms decide what millions of users see and that builds its advertising business around that distribution can hardly retreat into the role of a passive conduit when some of what it helps circulate is abhorrent. The Indian government’s contention that Meta should instead be regarded as a service provider gives it some rightly deserved...

Algorithmic Alibi

Meta’s handling of child sexual-abuse material has exposed the hollowness of the tech behemoth’s claim to be merely a simple intermediary. A company whose algorithms decide what millions of users see and that builds its advertising business around that distribution can hardly retreat into the role of a passive conduit when some of what it helps circulate is abhorrent. The Indian government’s contention that Meta should instead be regarded as a service provider gives it some rightly deserved raps on the knuckle. Meta has now agreed to report child-safety matters directly to the cybercrime portal managed by the Indian Cyber Crime Coordination Centre (I4C). Until now, such cases were reported to the United States-based National Center for Missing and Exploited Children. The new arrangement raises the obvious question of why a company with Meta’s technological resources required regulatory pressure before establishing a direct reporting mechanism with Indian authorities. Meta’s own business model has demolished its fiction of passivity. Facebook and Instagram do not simply sit there waiting for users to post material. Their algorithms rank, recommend, personalise and amplify that material. They determine what attracts attention and, by extension, what generates engagement and advertising revenue. Meta has invested enormous resources in making those systems increasingly effective. It cannot then plead helplessness when the same machinery distributes material that it would rather not acknowledge. Meta’s systems are notoriously difficult for outsiders to scrutinise, while the company possesses vast amounts of information about how they operate. Governments, researchers and users are therefore often expected to identify problems without possessing anything like Meta’s visibility into its own platforms. The company cannot reasonably demand trust while keeping so much of the machinery that determines online visibility behind corporate walls. That said, the government has acknowledged that the courts must ultimately determine whether the company retains safe-harbour protection. But legal immunity and corporate responsibility are not the same thing. The more sophisticated Meta’s technology becomes, the weaker its excuse for failing to anticipate predictable abuses becomes. A company capable of identifying a user’s interests with astonishing precision should not find child-abuse material an unknowable phenomenon. There is a revealing asymmetry in Meta’s posture. Algorithmic power is a selling point when it increases engagement. It becomes someone else’s problem when that power produces consequences that attract regulatory scrutiny. The company wants credit for knowing what people want, while retaining the legal convenience of claiming that it cannot know what appears before them. This is selective accountability. While the courts will decide the precise limits of safe harbour, Meta should not confuse legal uncertainty with moral ambiguity. A company that designs the machinery, controls its settings and profits from its operation cannot forever wash its hands of what that machinery helps to put in front of the public.

Ghibli vs Generative AI: A Battle for the Soul of Art

Dr. Sajid Sheikh & K. Ankita Rao
May 14, 2025
4 min read

As Studio Ghibli’s iconic style gets swept into the AI age, a storm brews over copyright, consent, and the soul of art.

Ghibli-inspired art has taken the internet by storm in recent weeks, especially the social media, where netizens are posting pictures converted into Ghibli Style images by Artificial Intelligence (AI). While the people seem to thoroughly enjoy the new feature of OpenAI, it has opened the doors for several debates, especially in the legal fraternity. It raises several questions regarding privacy, ethical concerns and most importantly about copyright infringement. The founder of Studio Ghibli, Artist Hayao Miyazaki himself expressed his displeasure over the Ghibli styles images of people doing circles on social media, calling the art generated by AI “an insult to life itself”.


The issue began with OpenAI launching GPT-4o which enabled users to upload pictures and turn those into unique high-quality images resembling distinct artistic styles of Studio Ghibli. While this brings out warm nostalgia and excitement, it has also brought to the forefront significant concerns about the legal limits of AI-generated art.


Studio Ghibli has a very unique style of art and an extraordinary way of connecting with people. It’s not just visually stunning, but also deeply emotional. With its detailed hand-drawn animation, it brings both magical moments and everyday life to the screen with equal care and attention. The soft, warm colours add a sense of comfort and nostalgia, making the scenes feel timeless. What truly makes Ghibli special is how effortlessly it blends fantasy into ordinary life. It’s this warmth and authenticity that makes the art feel not just beautiful, but truly human. The animations are carried out by skilled animators and therefore the replication of this style of art has become controversial in the recent times.


A key question is whether Studio Ghibli’s art style can be protected under copyright law. International frameworks like the Berne Convention and TRIPS safeguard only the expression of ideas, not the ideas themselves. Artistic styles, such as Ghibli’s distinctive aesthetics, are typically deemed unprotectable ideas; only specific characters and images qualify for copyright. Granting exclusive rights over a visual style risks stifling creative freedom. OpenAI argues that it uses only publicly available ideas to render images in a Ghibli-like manner. Yet courts apply various tests such as ‘substantial similarity’ and ‘look and feel’ to assess whether imitation crosses into infringement.


Courts often use the ‘look and feel’ and ‘substantial similarity’ tests to determine copyright infringement. The former asks whether an average observer finds the overall impression of two works strikingly similar, while the latter examines whether protectable elements like unique expression or structure have been unfairly echoed in the new work.


Though OpenAI’s outputs may resemble Studio Ghibli’s aesthetic, legal protection applies only to concrete, original expressions, not abstract artistic styles. Courts typically filter out unprotectable elements before applying these tests, making it difficult for Studio Ghibli to argue that its art style, however distinctive, qualifies as a protectable expression rather than a mere idea.


A derivative work, generally is one that is based on, adapted from, or recast of an existing copyrighted work. Under the Indian law, it is considered as an adaptation which is also a subject matter of copyright protection. Therefore, if an AI creates an image that looks a lot like something from Studio Ghibli and especially if that image was generated using copyrighted material as part of its training data without permission, it could be considered an unauthorized derivative work amounting to copyright infringement. Interestingly, if AI is simply inspired by a general artistic style, without copying specific characters, scenes, or unique design elements, it might not qualify as infringement. The debate again boils down to the question of whether the ‘art form’ in itself could be considered as an expression or a mere idea, which will determine its copyrightability.


A critical point of contention is whether using copyrighted material to train AI qualifies as fair use. Proponents argue that AI transforms ideas into new expressions based on user prompts, akin to how original works build on public knowledge. Critics, however, contend that even storing such material for training infringes copyright, as this right rests solely with the creator. In India, the stricter doctrine of ‘fair dealing’ further complicates the issue, raising doubts over whether AI training practices can be legally justified under existing frameworks.

As users upload personal photos to generate Ghibli-style images via AI platforms, concerns have mounted over whether such data is quietly repurposed for model training, raising red flags over privacy and potential misuse. If reused, AI could inadvertently replicate individuals’ likenesses in new artworks, blurring the line between consent and exploitation. Studio Ghibli’s co-founder, Hayao Miyazaki, once called AI animation “an insult to life itself” - a sentiment echoed by traditional artists who fear that hand-drawn craft, rich in emotion and cultural depth, is being undermined. Though not strictly a question of intellectual property, the ethical implications are profound.


The rise of AI-generated art in the style of Studio Ghibli has ignited complex debates around originality, authorship, and copyright. While such technology challenges existing frameworks, it also forces a re-evaluation of transformation, fair use and inspiration in the digital age. Striking a balance between protecting artists and fostering innovation is crucial. To safeguard creativity and progress, clear legal and ethical guidelines must follow to ensure that the future of art remains rich, respectful and imaginative.


(Dr. Sajid Sheikh is Assistant Professor (Law) and Ms. K. Ankita Rao is Research Assistant, DPIIT IPR Chair. Both are at the Maharashtra National Law University, Mumbai. Views personal.)

Comments

bottom of page