The concept of a patent in Australia is historic and can be traced back to feudal England during the reign of Queen Elizabeth I. The function of a patent largely remains the same today, providing a temporary monopoly over an invention. However, recent technological advancements in artificial intelligence (AI), software, and digital algorithms have required the law to adapt rapidly to these emerging technologies. This article explores the historic evolution of patents, and analyses how patents interact with new emerging technologies.

Evolution of the patent system

The common law traditionally approaches monopolies with caution. In contrast to a monopoly, free competition allows the price and quality of goods to adapt with evolving economic and social conditions. This encourages efficiency and flexibility within a market economy. Nonetheless, monopolies have been a consistent feature of market economies since their development in early Europe.

During the reign of Queen Elizabeth I in the United Kingdom (UK), the Crown granted monopolies to individuals who introduced new inventions to the realm. These grants were an exercise of the Crown’s prerogative power and were known as letters patent. Similarly, the Crown granted foreign trade monopolies, resulting in a handful of very large and powerful companies. For example, at its peak, the East India Company was larger than several nations. The East India Company’s exclusive overseas trading rights with India, China, Persia, and Indonesia provided affordable tea, cotton textiles, jewellery, and spices in England. The exchange of annual payments (known as royalties) in return for such monopolies meant that this was a lucrative practice for the Crown.

By the end of Queen Elizabeth I’s reign, many monopolies had been granted under the prerogative power. This caused widespread constitutional concern. In the 1602 case of Darcy v Allen,¹ monopolies were held to be generally invalid because they resulted in higher product prices and lower product quality. However, the Court allowed an exception to this rule. Inventions that introduced new trade into the realm remained eligible for a temporary monopoly. The subsequent Statute of Monopolies 1624 (UK) implemented this rule. Under the statute, valid patents for eligible inventions lasted for 14 years.

Due to the increasing number of patents granted under the Crown’s prerogative power, it became necessary for inventors to provide specifications along with their application. This involved a technical document explaining the invention and how it could be put into practice. Specifications not only allowed an invention’s eligibility for a patent to be assessed, but also allowed an expert to perform the invention at the end of the patent term. This requirement recognised the expectation that inventors should eventually pass on their knowledge for the benefit of the realm.

Comprehensive legislation was introduced in Britain in the 20th century to streamline and simplify the process for obtaining and enforcing patent rights.² This served as the model for Australian legislation, resulting in the Patents Act 1952 (Cth). The Patents Act 1952 (Cth) established a Register of Patents under which patents could be obtained and registered for a 16-year term. Patent applications were processed by the Patents Office and led by the Commissioner of Patents. This system has largely remained in place under the current Patents Act 1990 (Cth), with few substantive changes. Under the Patents Act 1990 (Cth), a standard patent can last for up to 20 years, and a pharmaceutical patent can last for up to 25 years.

Today, patents remain an invaluable source of protection for new inventions against market competitors. A patent confers exclusive legal rights to exploit the invention and authorise others to exploit the invention.³ These rights are considered personal property and are capable of assignment and devolution by law.⁴

Patent law and emerging technologies

The contemporary advent of AI and machine learning has resulted in an explosion of technological inventions. These emerging technologies are increasingly complex and raise issues of patentability and inventorship. The law has had to adapt rapidly in response to accommodate the use of software, digital algorithms, and AI.

Is AI an inventor?

Certain AI programs have the capability to autonomously produce inventions that would ordinarily be eligible for patent protection. This raises the issue of whether AI can be an inventor for the purpose of obtaining a patent under the Patents Act 1990 (Cth). Section 15(1) of the Patents Act 1990 (Cth) provides that a patent may only be granted to a person who:

1. is the inventor;
2. would be entitled to have the patent assigned to them;
3. derives title to the invention from the inventor or a person mentioned in paragraph; or
4. is the legal representative of a deceased person mentioned above.

In Commissioner of Patents v Thaler,⁵ Dr. Stephen Thaler applied for a patent for a ‘food container and devices and methods for attracting enhanced attention’.⁶ In the patent application, Dr. Thaler claimed that the invention was autonomously generated by artificial intelligence. Dr. Thaler did not claim that he was the inventor. Consequently, the Full Court of the Federal Court was required to determine whether an AI machine could be considered an inventor under s 15(1)(a) the Patents Act 1990 (Cth).

The Court held that an invention devised by an AI system is not patentable. The Court further observed that the grant of a patent under the Patents Act 1990 (Cth) is based upon an invention that arises from the mind of a natural person. Essentially, a human must author a patentable invention.

This outcome emphasises the origins of the patent system and reinforces that the grant of a limited term monopoly is a reward for human endeavour. It is also consistent with other common law jurisdictions. In the United States case of Thaler v. Vidal,⁷ it was held that AI cannot be an inventor for the purposes of the US patent system. In the context of the UK patent system, the same conclusion was reached in Thaler v Comptroller-General of Patents, Designs and Trademarks.⁸

Are algorithms patentable?

A mathematical algorithm is a finite sequence of rules or instructions that is processed by a computer to produce an output. Algorithms are useful for solving complex problems at high speed, and unique algorithms can be a valuable asset of a business.

However, algorithms are considered a type of ‘scheme’ or ‘abstract idea’ which is not patentable at Australian law. In Grant v Commissioner of Patents,⁹ the Full Court of the Federal Court held that intellectual information such as a mathematical algorithm, working directions, or business scheme is not patentable because it does not produce a physical effect.

Although an algorithm itself is not patentable, an algorithm that is part of an otherwise patentable invention will not exclude a claim from patentability. A distinction is drawn between a claim to an abstract algorithm and a process that produces a material advantage or use.

Nonetheless, abstract algorithms can be valuable tools that a business may wish to protect. For example, Google’s search engine utilises an algorithm to fulfil complex web searches at high speeds. This algorithm is essential to the search engine’s function and quality of user-experience. The superior function of the algorithm allows Google to distinguish themselves from market competitors.

Such algorithms may be protected as confidential information or a trade secret. Confidential information confers the right to prevent other people from using or disclosing the algorithm. However, unlike patents, confidential information is not protected by registration or legislation. Confidential information is protected at common law and equity,¹⁰ and may be contractually enforced by private agreement between parties.

Are computer-implemented inventions patentable?

A computer-implemented invention (CII) involves the use of a computer, computer network, or other programmable device. At least one feature of the invention must involve a computer program or other software. This requirement distinguishes CIIs from inventions that merely use computers as a tool to achieve an existing business method or scheme. A CII must relate to the specific way in which the computer carries out the scheme or method, and should exceed the ‘known functions’ of the computer.¹¹ Known functions could include data storage, display, data processing, and basic GPS operation (depending on the invention).¹²

The Patents Act 1990 (Cth) does not directly exclude CIIs or software. Under s 18 of the Patents Act 1990 (Cth), an invention will be patentable if, amongst other things, the invention is a ‘manner of manufacture’ within the meaning of s 6 of the Statute of Monopolies 1624 (UK). This effectively requires the patent to be directed towards an appropriate subject matter. In the seminal 1959 case of National Research Development Corporation v Commissioner of Patents (NRDC)¹³, it was held that an invention will generally be a manner of manufacture if it produces an artificially created state of affairs that is useful. However, the common law has struggled to adapt this test to CIIs.

In contrast, the patentability of CIIs in other major common law jurisdictions is relatively more settled. For example, in the UK, a CII is patentable if it provides a technical contribution.¹⁴ Similarly, a patentable CII in Canada must exhibit a discernible effect or technical change¹⁵. In the United States, a CII must provide a technological improvement of a computer’s functionality.¹⁶

The Aristocrat litigation provides a recent example of the Australian common law grappling with adapting the traditional test to CIIs. In 2016, Aristocrat (one of the largest gaming services providers in the world) applied to the Australian Patent Office for patents relating to their electronic gaming machines. Specifically, Aristocrat’s application concerned a system and method for providing a feature game. Their slot machines programmed symbol sequences on virtual reels to trigger these bonus feature games. The application was rejected by the Australian Patent Office.

The matter was subsequently appealed multiple times. The Full Court of the Federal Court initially proposed a two-stage test to determine if a CII is patentable. The matter was then appealed to the High Court, however, the Court was split 3:3 as to whether the alleged CII was patentable. Importantly, the High Court was unanimous in its rejection of the two-stage test and remitted the matter back to the Full Court of the Federal Court.

In the latest decision of Aristocrat Technologies Australia Pty Ltd v Commissioner of Patents,¹⁷ it was noted that the High Court’s rejection of the two-stage test was a compelling reason not to continue with the two-stage test. Instead, the Full Court of the Federal Court reverted to the traditional principles established under NRDC, ultimately holding that CIIs are patentable.

However, the law is not yet settled. The Commissioner of Patents recently made a special leave application, arguing that the Full Court erred in finding that the invention is a manner of manufacture within the meaning of s 18(1A)(a) of the Patents Act 1990 (Cth). Accordingly, it remains unclear what the exact formulation of the test for the patentability of CIIs is.

Next steps

The concept and system of patents in Australia is largely inherited from the UK and can be traced back to feudal England. However, patents remain a invaluable tool for protecting inventions against market competitors. Patents also interact with AI, software, and algorithms in unique ways. Whilst the specific test for patenting a CII remains uncertain, it is clear that AI cannot be an inventor for the purposes of a patent. Additionally, an abstract algorithm may not be patented on its own, but can be protected in other ways.

HWLE Lawyers’ Intellectual Property team has extensive experience advising on the patentability of inventions involving emerging technologies. If you have any questions or concerns regarding artificial intelligence and copyright infringement, please do not hesitate to contact us.

This article was written by Luke Dale, Partner.

¹ (1602) 74 ER 1131.
² Patents Act 1949 (UK).
³ Patents Act 1990 (Cth) ss 15(1)(a), (c).
⁴ Ibid s 13(2).
⁵ [2022] FCAFC 62.
⁶ Ibid.
⁷ 43 F.4th 1207, 1210 (Fed. Cir. 2022).
⁸ [2023] UKSC 49.
⁹ (2006) FCAFC 120.
¹⁰ Coco v AN Clark (Engineers) Ltd [1969] RPC 41.
¹¹ ‘What Computer-Related Inventions Can Be Patented?’, IP Australia (Web Page) <https://www.ipaustralia.gov.au/patents/what-are-patents/what-computer-related-inventions-can-be-patented#:~:text=When%20it%20comes%20to%20computer,Known%20functions%20can%20include:>.
¹² Ibid.
¹³ (1959) 102 CLR 252.
¹⁴ Lenovo (Singapore) PTE Ltd v Comptroller General of Patents [2020] EWHC 1706 (Pat).
¹⁵ Victor Stanley Inc., Re. 2012 CarswellNat 885.
¹⁶ Enfish, LLC v. Microsoft Corp., 56 F. Supp. 3d 1167 (C.D. Cal. 2014).
¹⁷ [2025] FCAFC 131.