In our previous article, we explored how to prove a work is yours. This week, a harder problem: what happens to that ownership when artificial intelligence helped create the work in the first place? It is a question Malaysian law has not yet fully answered, and the gap matters for anyone building a business on original content.
Copyright has always rested on a simple assumption: that a human being created the work. The Copyright Act 1987 is built around human authorship, a work reduced to material form by a qualified person. Generative AI unsettles that foundation, because AI is not a legal person and cannot, on its own, be an author. That leaves two live questions.
Is an AI-assisted work even original?
To be protected, a work must be original. Under Section 7(3) of the Copyright Act 1987, that means sufficient effort was expended to create it. In YKL Engineering v Sungei Kahang Palm Oil [2022], the Federal Court tied this to skill, time and labour expended by a human author.
So the real question becomes: how much of the work was human? Different jurisdictions answer differently:
- The United States takes a strict line. In the Zarya of the Dawn decision and Thaler v Perlmutter (2023), courts held that human authorship is the bedrock of copyright, and that even hundreds of detailed prompts do not turn AI output into human-authored work.
- China is more flexible. In cases such as Li v Liu (2023), courts recognised copyright where a human made meaningful choices in selecting, arranging and designing the AI output, treating AI as a tool rather than the creator.
Malaysia has no direct case yet. The stronger view is that a work generated by AI with minimal human input may fail the originality test, while a work involving genuine human skill and judgment may still qualify.
The safer your human fingerprints on the work, the safer your copyright. Bare prompts are the weakest ground.
Who is the author when AI is involved?
Under Section 10 of the Act, the author must be a qualified person, defined in Section 3 as a citizen or permanent resident, or a body corporate established in Malaysia. AI is neither, so under the current framework AI cannot be an author or first owner of copyright.
The wider structure of the Act points the same way. Provisions that tie the term of copyright to the author's lifespan, and that grant moral rights, only make sense for a human creator. Notably, unlike the UK's Copyright, Designs and Patents Act 1988, Malaysia has no specific provision for computer-generated works, which may mean purely machine-made works fall outside protection altogether. The much-discussed "monkey selfie" dispute captured the same core point: copyright law expects a human hand behind the work.
Why this is not just theory
These questions are already reaching the market. The controversy over the novel Shy Girl, where a large share of the text was alleged to be AI-generated, shows how quickly disputes over originality, authorship and responsibility can surface, and how hard it is becoming to tell human creativity from machine output. For a business, the risk is concrete: content you assume you own may be weaker than you think, and your ability to stop copycats depends on it.
What creators and businesses should do now
Until Malaysian law catches up, protect yourself with good habits rather than assumptions:
- Keep the human in the loop, and keep the receipts. Record the creative choices, edits and direction your team contributed, not just the prompts.
- Do not rely on pure AI output for anything commercially important without meaningful human authorship layered on top.
- Check your contracts. Make sure agreements with staff, freelancers and agencies address AI use and assign whatever rights do exist to you in writing.
From a recent talk
In March 2026, our IP team spoke to the Writers Society on exactly this frontier, from how copyright arises to what AI now means for originality and authorship. The clear takeaway: the law is still catching up, so the creators best protected are those who can show real human craft behind their work.
