IP Insights

One of the biggest misconceptions surrounding artificial intelligence is that open AI models and patents are fundamentally opposed. If powerful AI models become openly available, some assume patents will become less relevant because the technology is being shared rather than protected. At first glance, that sounds logical. In reality, the opposite may be true.

Open AI models and patents serve different purposes. Open models accelerate the spread of knowledge, reduce barriers to innovation and allow more people to experiment with emerging technologies. Patents, on the other hand, encourage investment by rewarding genuine technological advances. Rather than competing with one another, they can strengthen the same innovation ecosystem from different directions.

NVIDIA's recently published policy paper on frontier AI reflects this balanced approach. Rather than suggesting that the future belongs exclusively to either frontier closed models or frontier open models, NVIDIA argues that both have an important role. Frontier closed models continue to justify the enormous investment required to push the boundaries of AI capability. Frontier open models broaden access to that capability, encouraging scientific research, cybersecurity, education, competition and downstream innovation by allowing others to build upon those foundations. While the paper focuses on the future of AI, it also raises an interesting question from an intellecual property perspective. If wider access to advanced AI capabilities leads to more innovation, could it also lead to more inventions deserving patent protection?

This is not the first time a transformative technology has raised questions about the future of innovation and intellectual property. 

At the beginning of the Industrial Revolution, it would have been easy to assume that as machines became more widespread and manufacturing techniques spread across industries, invention would become commoditised and patents would gradually lose their importance. Instead, the opposite happened. The Industrial Revolution accelerated technological progress, created entirely new industries and generated countless new engineering challenges. Greater access to technology did not reduce innovation. It created more opportunities for innovation.

A similar concern could have been raised when the internet became widely accessible. Information could suddenly be shared around the world almost instantly. Some questioned whether intellectual property would become less valuable in an age where knowledge flowed so freely. Instead, the internet became the foundation for entirely new technologies, industries and business models. From cloud computing and e-commerce to cybersecurity and artificial intelligence itself, innovation expanded rather than contracted.

AI models, especially, frontier open AI models, may represent the next stage of that same pattern. As access to powerful AI models becomes more widespread, the models themselves become less of a competitive advantage. Instead, value shifts to what businesses, researchers and inventors build with them. The real innovation increasingly lies in improving performance, reducing computational cost, integrating AI with hardware, developing new technical applications, solving engineering problems or creating entirely new products and services. Those are precisely the types of advances that patent systems have traditionally been designed to encourage.

The wider availability of frontier open models may also improve patent quality. Every published model, research paper and technical disclosure adds to the body of prior art. That makes it more difficult to obtain broad patents over concepts that are already publicly known, encouraging applicants to focus on genuine technological advances that are both new and inventive. Greater openness does not necessarily weaken the patent system. It may strengthen it by raising the threshold for what deserves protection.

This broader principle is also reflected in the patent laws of Australia and New Zealand, where the focus remains on identifying a genuine technological contribution rather than simply relying on the use of software or AI. Neither jurisdiction is likely to reward an invention simply because it incorporates AI. In New Zealand, section 11 of the Patents Act excludes computer programs as such, requiring the focus to remain on the substance of the invention rather than software alone. In Australia, computer implemented inventions must still satisfy the established requirements for patentable subject matter. The concept of a useful, artificially created state of affairs continues to form an important part of the broader manner of manufacture inquiry. As frontier AI models become increasingly accessible, the key question in both jurisdictions is unlikely to be whether AI was used, but whether the claimed invention provides a genuine technological contribution beyond simply using the underlying AI model as a tool.

The relationship between closed models, open models and patents should not be viewed as contradictory. Closed models continue pushing the technological frontier through significant investment in research and infrastructure. Open models broaden participation by allowing a much larger community of businesses, researchers and inventors to build upon those advances. Patents, in turn, continue encouraing investments in the genuine technological advances that emerge from both. Each performs  different but complimentary role within a healthy innovation ecosystem.

The debate, therefore, should not be framed as open versus closed or open models versus patents. History suggests that when transformative technologies become more accessible, innovation does not diminish. It accelerates. There is every reason to believe artificial intelligence will follow the same path. Frontier open AI models are unlikely to reduce the importance of patents. Instead, they may make genuine technological innovation more valuable than ever.

At IPZEAL, developments like these are followed closely because they provide valuable insight into where innovation and intellectual property are heading. As AI continues to evolve, understanding the relationship between openness, investment and patent protection will become increasingly important for innovators across Australia, New Zealand and beyond.

About the Author

Pritesh Lohani is a Registered Patent Attorney in Australia and Registered Trade Mark Attorney in Australia. He is the founder of IPZEAL and has over 16 years' experience advicing clients on patents, trade marks and designs across Australia, New Zealand and internationally.