Singapore strengthens IP framework amid AI rise

Caption language: en · Fetched: 2026-08-27

Singapore plans to examine its intellectual property regimes to keep pace with the rise of artificial intelligence. It is seeking public views on key copyright and patent issues, including who should be recognized as the creator of AI assisted work. Law Minister Edwin Tong also urges firms to make intangible assets a core part of their business strategy. Nividi has more. There's more to this cup than just coffee. The process behind brewing this drink is a closely guarded secret. For this local startup, protecting its intellectual property goes beyond how its coffee is made. Even the data generated by its vending machine can be deemed as a valuable intangible asset. >> We have currently more than 170 machines islandwide in Singapore. So and every of these machine is generating telemetry data.

So sales data it is telling us whether the machine has errors, has fault, technical faults, right? So all of this data uh goes back and help us with our operations uh our uh maintenance right the technical work as well as our uh sales aspect of our business. At the opening of Singapore IP week, law minister at Winong says intangible assets like software and data are increasingly key to staying competitive. He added that artificial intelligence has dramatically changed where businesses derive their value. >> If anyone can generate a good advertisement in seconds, the brand behind it may then matter more than the advertisements around it. If thousands of product designs can be generated overnight, the advantage may less may lie less in generating them and more in knowing with data what your consumers want.

Know how and networks then to bring it to the market. >> And that's why Singapore plans to strengthen its intellectual property framework. It's to help businesses here and abroad find value in its assets through new guidelines and programs. It is also seeking public views on how AI affects patent and copyright issues. For example, whether copyrighted works can legally be used to train AI. When content becomes abundant, true human ingenuity becomes more, not less invaluable. IP will increasingly be about not just protecting what is created, but also establishing who created it. And I think this is a new question or a new dimension that we should think about carefully. >> Singapore will also deepen its international partnerships to develop a common language around intangible value.

This could make it easier for these assets to be invested in and transacted across borders. >> The World Intellectual Property Organization plans to step up efforts to make intellectual property more widely understood and recognized. It says IP should no longer be seen as a technical legal matter. The agency also wants to play a bigger role in supporting global dialogue to address complex challenges. IP must continue being a way for us to collaborate and cooperate even in a world where political divide and conflict is becoming more common. I've often said that invention is a solo endeavor but IP is a team sport. And almost every device that we use. Let's take the smartphone as an example embidates innovations and inventions that cut across hundreds if not thousands of enterprises, inventors and innovators.

For more on how AI will make an impact on copy vikes and patents, we speak to chief executive of the intellectual property office of Singapore, Tan Kong Hwi. Uh Kongwi, let's begin by asking this question. If AI is used in coming up with an invention, uh, who actually gets to call themselves the inventor and looking at our current patent rules, does it still apply in situations like this? >> That is a great question, Paul. Uh, and it's many such questions have appeared in the world since generative AI was introduced about 3 years ago.

uh so that is why we made this year's uh Singapore IP week uh uh h have a theme around AI and uh you know the conference started today uh it had more than,500 participants and uh coming from more than 50 countries now going back to your question uh around uh inventorship uh and who are the creators uh so our regime is one that recognizes rewards and incentivizes human creativity and inventiveness Uh so uh although AI is a very sophisticated tool for creating things for inventing things uh today to get protection we still require a human creator as well as a human inventor. >> Now we all know that AI can turn out huge amounts of technical information. Could that make it harder to decide uh whether invention is actually new? >> Yeah, another great question.

So uh whether a piece of information is created by AI or whether that is created by a human being actually in the course of patent examination it doesn't really matter our patent examiners will look at the relevance of that piece of information uh in assessing whether a patent can be granted or not right uh and you know uh IPOS has the hallmark for producing quality examinations and we hope to continue to uphold that quality now the real issue is with AI generating uh a lot of information that amount of work to look through that information increases and that can drag down the efficiency of our IP offices. Uh so in the same way that technology has helped us generate more information, I think it can also aid us uh in producing better search tools, better tools for analysis.

Uh so I think in a sense uh I think in a very near future uh this will not be a problem. Now to understand it this clearer, if I give AI say a prompt and then refine what it produces, at what point can I call that work mine and and how much of the creative input needs to come from me for copyright to be applied here? >> Yeah. So the work can potentially be called yours. Uh but it does depend on whether there is uh sufficient engagement of the human intellect. Um so let me give you an example here.

Uh so if uh if a piece of work you've just given it one or two prompts and you know it's mostly generated by AI that is unlikely to be protected and it probably cannot be called yours right uh but if you have uh done a lot of work to shape the output to shape the outcome of the work given a lot of creative inputs uh to the AI engine uh the chances for protection is higher. Now ultimately whether it meets the measure of having crossed that line for protection that is for our courts to uh assess. Uh we have engaged with many creators uh some of it is through the visual audio creative content professional association and we know that creators are struggling with this exact uh question and this is why IPOS has published uh AI and IP explainers. Uh it's on our website to help people understand.

We give them practical tips around how they can document uh the human contribution in their work. Uh so that when it comes to the crunch, they have proof that you know it was a human that contributed sufficiently to uh the piece of work. >> Now speaking of creatives, a big concern for creatives is uh those who use their work to train AI. Um how do we protect their rights without making it difficult or costly for AI companies to innovate? Yeah, I'm glad to say that the Singapore regime today uh has carefully balanced uh the needs uh and of uh both stakeholders groups, right? Uh on one hand uh the innovators, the tech companies uh they today have legal clarity on uh how they can use copyrighted works for purposes of tax and data mining.

Uh on the other hand uh our laws uh also protect the interest of our copyright holders right and we do that through safeguards. Uh an example of this safeguard would be that uh the innovators will have to lawfully access uh the the data or the copyrighted works and we do this through a piece of uh exception in our copyright law and it's called the computational data analysis exception. Uh so with this exception actually it gives a lot of clarity to both sides uh and it balances the interests of both parties. Now for the everyday users of AI tools um like myself uh how do we know that if we're accidentally infringing on uh someone else's copyright >> what can people in business perhaps do to protect themselves?

Yeah, I think the first step is what exactly what you have done which is having some awareness that uh you know the these AI tools while they are powerful uh you can be more prone to infringing copyright uh when when you use these tools to generate images and works. Uh so then the next is about taking some practical steps and that is really about checking the output uh that you get from uh these AI uh applications uh and then uh making sure that they have not uh infringed on some some other works. So there are in fact already many digital tools available to do that. Uh one of them would be uh a reverse image uh search tool. You can use a reverse image search tool to verify if there are similarities with existing works.

uh and I think this is a really important step to take especially uh if you are trying to use these uh AI generated output for commercial purposes. Now a quick question our last question actually um looking at the bigger picture. How does Singapore need to get right if it wants to be a trusted IP hub in the AI era? >> I think we are already on track on becoming a leading IP hub uh for the AI economy.

uh we have many strong uh fundamentals uh today you know the in fact the value propositions that make us a leading hub today uh which is trust quality efficient examination uh a very capable uh IP services sector that can serve the most sophisticated needs of MNC's and companies I think these are value propositions that remain valid uh going into the future as well so I have a lot of confidence that we can remain uh a trusted IP hub uh in the AI economy going in are looking to the future. >> Well, thank you so much for sharing your insight with us on this matter. We've been speaking to the chief executive of IPOS, Tan Kongwi.