Singapore has opened the consultation that will determine how its intellectual property regime handles artificial intelligence. The Ministry of Law and the Intellectual Property Office of Singapore launched it on 26 August 2026, and submissions close at 5pm on 22 October 2026.

The copyright questions

The consultation paper addresses three areas on the copyright side.

The first is certainty and accountability in AI training. This is the text and data mining question that has produced litigation in the United States and legislative activity in the European Union and United Kingdom: on what basis may copyright works be used to train models, and who answers for it when they are used without permission.

The second is copyright risk management in deployment and use. This is the downstream problem of what happens when a model reproduces protected expression in its output, and how deployers rather than developers should manage that exposure.

The third is the nature of human creativity in AI-assisted works, which asks how much human input a work requires before it attracts protection at all.

The patent questions

Two areas are raised. The first is inventorship: how existing principles apply across a widening spectrum of human-AI interaction in the inventive process. Singapore has already been examining adjacent ground, having consulted earlier in 2026 on the implications of the UK Supreme Court’s Emotional Perception decision concerning artificial neural networks and patentable subject matter.

The second is more novel and has attracted less attention than it deserves. Large-scale publication of AI-generated technical disclosures may alter the prior art landscape. If it becomes trivial to generate and publish plausible technical descriptions across an entire field, the pool of material capable of destroying novelty expands enormously, potentially as a deliberate defensive strategy. Any organisation maintaining a patent portfolio has an interest in how Singapore treats that.

Why respond

Consultation responses shape outcomes, and this one will decide whether Singapore aligns closer to the permissive commercial data mining approach, the European opt-out model, or something of its own. Given Singapore’s position as a regional headquarters location and its ambitions in AI development, the answer will influence where model training is conducted in Asia.

Organisations with material exposure, including content owners, model developers, deployers integrating third party models and patent-heavy businesses, have until 22 October. The consultation paper is available at go.gov.sg/ai-ip2026, with submissions via FormSG or email to the Ministry of Law.

Summary

MinLaw and IPOS are consulting until 22 October 2026 on how AI should reshape Singapore’s IP regime, covering AI training and accountability, deployment risk, human creativity in AI-assisted works, inventorship across human-AI collaboration, and the effect of mass AI-generated disclosures on prior art. The outcome will determine Singapore’s position relative to the UK, EU and US on text and data mining, and will influence where AI development happens in the region. The window for input is open now.