In the past few weeks, social media in Malaysia has been buzzing with a fascinating trend. People have been turning their Raya photos into Studio Ghibli-style artwork using AI tools. Suddenly, feeds were filled with dreamy, whimsical scenes that looked like they were pulled straight from Spirited Away or My Neighbor Totoro, but with familiar Malaysian faces and festive baju Raya. The results were undeniably beautiful and sparked a wave of excitement. Many users felt a deep sense of nostalgia, while others simply enjoyed the fun of seeing themselves in a fantasy world.
But beneath the surface of this creativity lies a legal and ethical concern that deserves more attention, especially here in Malaysia.
Studio Ghibli’s art is not just a style. It is the result of decades of work by talented animators led by the legendary Hayao Miyazaki. Their films are known for their distinct look, strong visual storytelling, and emotional depth. So when AI platforms start creating images that imitate this signature aesthetic, often trained on existing works without permission, it raises important questions. Are we witnessing innovation at its best, or are we stepping into a space that might involve copyright infringement?
In the United States and Europe, lawsuits have already begun. Several artists have come forward to challenge the use of their artworks in AI model training. The core concern is that these models may not simply take inspiration, but may instead reproduce recognizable features and patterns that go beyond what copyright law permits. There are also wider questions about whether using copyrighted works to train AI systems, without the consent or knowledge of the creators, should be considered acceptable in the first place.
While these developments are unfolding abroad, they raise equally relevant issues here in Malaysia. Under our Copyright Act 1987, artistic works such as illustrations, characters, and animations are protected automatically upon creation. If a copyrighted work is used to train an AI model, whether intentionally or otherwise, and that model later produces outputs that closely resemble the original, there may be grounds to claim reproduction or adaptation. Both of these are protected rights under our law.
Of course, the situation is not always straightforward. One of the central principles of copyright law is the idea-expression dichotomy. Copyright protects the expression of an idea, not the idea itself. In practice, this means that while a general artistic style like Ghibli’s may not receive copyright protection on its own, specific visual elements such as characters, compositions, or iconic scenes may very well be protected. This is where the line between inspiration and imitation becomes difficult to define.
Another layer of complexity involves how AI models are trained. Many developers rely on large datasets collected from the internet, which may contain copyrighted works. In response to criticism, some AI companies have argued that this kind of use qualifies as fair use, or what we call fair dealing in Malaysia. They claim that the outputs are transformative and therefore do not infringe the original work.
However, Malaysian copyright law provides a much narrower exception for fair dealing compared to the broader fair use doctrine in the United States. Here, fair dealing applies only in limited situations such as private study, research, criticism, review, or news reporting. It does not extend to the commercial training of AI models. Therefore, if copyrighted artworks were used to train a system that generates outputs for public use or commercial gain, and no permission or license was obtained, it could lead to serious legal implications under our current legal framework.
In contrast, some jurisdictions are beginning to address this issue more directly. The European Union has introduced policies that allow copyright holders to opt out of having their content used in data mining, effectively requiring AI developers to seek permission. Malaysia may one day follow this path, but for now, there remains a significant regulatory gap.
Beyond the legal questions, there are broader concerns about fairness and recognition. Many artists, both in Malaysia and internationally, have voiced frustration that their work is being used without credit, consent, or compensation. While AI tools can be useful and inspiring, the fear is that they may eventually diminish the value of original creative labour, especially when they are built on content created by others without proper acknowledgment.
This is not a rejection of AI or technology. AI can play an exciting role in the creative world. But it should be developed with consideration and respect. Just because a tool can recreate a Ghibli-style image does not mean it should be used freely, without asking who the original work belongs to and whether those creators deserve recognition or compensation.
Malaysia’s legal framework, like those in many countries, has not fully kept pace with the rise of generative AI. In the meantime, developers, content creators, and casual users alike should remain aware of how these tools are built and how the content they produce fits within our legal and ethical systems.
Moving forward, we need to find a balance between promoting innovation and protecting original work. This is not just a legal discussion. It is also a cultural one. Artists, educators, lawyers, developers, and users all have a role to play in shaping how we move through this new creative landscape.
I would love to hear what others think, especially those working in the creative or legal sectors. How do you see this trend? Are we celebrating a new era of artistic collaboration between humans and machines, or do we risk overlooking the people whose original work made it all possible?
