AI is dramatically changing the landscape of many industries and the legal industry is no exception. When it comes to patent registration, the Australian Federal Court recently reaffirmed that patents must list human inventors, not AI systems. This decision aligns with the earlier 2021 ruling in Thaler v Commissioner of Patents, where the court initially allowed an AI system, DABUS, to be named as an inventor. However, this was later overturned, and the High Court declined to hear an appeal, solidifying the stance that only natural persons can be recognised as inventors under Australian patent law.
Seeing how intellectual property itself is widely considered to be a category of property that includes intangible creations of the human intellect – the issue of AI being granted legal rights under intellectual property law is bound to be challenged over the years as it develops and evolves.
What is Australian Patent Law?
A patent is a legally enforceable right for a device, substance, method or process. For your application to be successful, your invention must be new, useful and inventive or innovative. When granted, a patent will give you exclusive commercial rights to your invention (a monopoly) with a standard patent lasting for up to 20 years.⠀
For information on Australian patent law and the patent application process, IP Australia is a fantastic resource. Comprehensive information on patents can be found here.
What is the Patentability of AI-Generated Inventions?
When it comes to patenting inventions created or assisted by AI, Australian patent law requires the invention to demonstrate a “technical effect” or solve a “technical problem” to be eligible. This requirement often excludes software-only innovations or AI-generated outputs that lack a tangible technological advancement.
And What About the Patentability of AI Technology Itself?
While AI systems cannot be named as inventors, AI technology itself can be patentable in Australia – but only in specific circumstances. To qualify for protection, the invention must produce a tangible technical effect or solve a specific technical problem. For example, a novel machine learning algorithm that improves processing speed, a unique neural network architecture, or an innovative AI-powered image recognition method could potentially be eligible. However, simply applying a standard AI model to an existing problem, or using AI for generic data analysis, is unlikely to meet the legal threshold.
AI Governance in Australia
Although Australia does not yet have a comprehensive legal framework dedicated solely to AI, the government is actively working on regulations. The Department of Industry, Science and Resources has proposed mandatory safety “guardrails” for AI use in high-risk contexts, alongside voluntary AI Safety Standards.
These regulatory efforts aim to promote responsible AI innovation while addressing potential risks, including those related to intellectual property and patent law.
