Software Patent: How Can It Be Done?

Even though life without computers is today nearly impossible to imagine, patent-based protection of the necessary software supporting such devices is still somewhat abstract. This is quite surprising, given that software-based instructions are necessary for operating everything from a modern fridge to most newly built vehicles. This article examines the usual hurdles and optimal approaches to patent software processes.

Can you obtain a software patent?

The intangible nature of computer programs allows them to be reproduced easily and in unlimited quantities, which complicates the way they can be controlled and protected. Copyright itself may be able to cover the source code of the relevant software, but not the actual idea of it, which is the core of its commercial value.

Therefore, the product-focused and exclusivity-granting patent protection, would appear to be a more realistic and dependable solution. This is also why, as seen in the diagram below, software patents and patents mentioning software have experienced an upward trend over the previous years.

Exponential growth of patenting involving the term “software”. Find out more here about our analytical software capabilities.

Moreover, since patents cover commercial ideas and products under a wider umbrella of exclusivity, more than just a single line of code or algorithm can be protected, meaning that a general market application and series of products can fall under the same patent’s protection of a certain software. This makes the entire process far more worthwhile for major companies.

What is required to patent software?

Several hurdles exist for obtaining a software patent. European Patent Convention statutes of the EPO exclude specific fields of knowledge, such as mathematical methods, methods for doing business and computer programs, which naturally impact on software on several levels.

Even so, it is vital to consider the “technicality” aspect, a European Patent Convention requirement in Article 52(1). This usually means that the software needs to be bound to actual technology, such as hardware, for example, a computer processor, to pass this first technical step. Essentially, the patent’s claims need to be “bound to hardware.”

Following that, the second hurdle involves the usual inventive step requirement, but with several interesting bespoke features. For example, the inventive step must also be of a technical nature and be associated with actively solving a certain problem or causing an improvement with an existing “technology”, such as an AI algorithm that devises new angles for antennae parts to increase their signal efficiency. In fact, such a computer-aided simulation of angles is purely mathematical, which should render it unpatentable. However, the essential enabler of the inventive step was the AI-based equation, which is, therefore patentable, since the related antennae is certainly a technical object.

This is mirrored in other European jurisdictions such as Germany, where, according to Section I(3) of the Patent Law, computer programs are expressly excluded from patentable subject matter. However, the very next legal section introduces that program-related inventions with a technical character can be patented, something that has already been confirmed by German courts in cases such as Federal Court of Justice, X ZR 110/13 (25 August 2015) ‘unlocking a screen of a mobile device’.

Major software-focused players in German, displayed by the numbers of patents and legal status

The USPTO, in the meantime, employs a slightly more flexible test, which has resulted in more open yet undefined software patent trends on the other side of the Atlantic. The so-called “Alice Test”, includes two surmountable hurdles for patenting. As part of the first step, it has to be decided if a patent claim is “directed to” a judicial exception to patentable subject matter, such as “abstract ideas”. Software, meaning the algorithm describing what the software invention does, in general is considered an abstract or “ineligible” idea that needs to be transformed into a “legible” one.

During the second step, only reached if the first step determines that the patents are directed to an abstract idea, we have to determine if there is  “something more” than the judicial exception that would make the claim patent eligible. This “something more” is an inventive concept that makes the invention patentable. As a rule, we have to find a practical application, like when we apply a software to effect a particular treatment or transform an article. In a sense it resonates a bit with the “technicality” aspect of the EPO.

Therefore, due to the added flexibility of the software patent tests in the US, software patents there represent far higher numbers, conflicts and value indicators when compared to the EP, where a higher level of originality is expected, and families are often larger in order to benefit from the cumulative protection of its members, as can be seen in the graph below.

Here we compare the different value indicators of a US and an EP software patent. Find out more here about such market evaluation projects.

Perhaps the optimal example of contrasts between the EPO and USPTO approaches’ can be gained by comparing different patents covering the same invention. A recent case in front of the EPO’s Enlarged Board of Appeal involved an algorithm through computer simulations that could improve electrical circuits, a process named “1/f noise”. The Board of Appeal reaffirmed the patent based on the technical aspect of “electric circuits” found in its claims.

However, if we examine the US equivalent patent US 6795840 B1, there is no mention of any electric circuits in the corresponding claim. Hence by the EPO definition, this patent would have been treated as purely mathematical, and therefore most probably quite vulnerable to invalidity proceedings.

Comparison of EP and US patent claims, with the absence of technical terms in US version

Moving to other jurisdictions, the requirements can be similar. According to the Japanese Patent Office in 2015, the core concept of “creating a technical idea utilizing the laws of nature” under Art. 2 (1), is met in the case of software, if said algorithm is attached with a hardware resource that works towards a specific purpose. Here, we can already see the similarities with the technological step of the EPO. This development of Japanese software patent law was made in 1993, when a new, also two-level, test was established.

The first ”utilizing the laws of nature” test, deserves an entirely separate article in of itself, but, highly summarised, is a more complex patentability test, which determines that actual “laws of nature” cannot be patented and that anything going against these laws also does not apply. This also covers certain “artificial arrangements” such as overly abstract ideas and methods of human mental activity, which also are unpatentable. Once that test is passed, the claim needs to clearly describe inventions that feature the utilization of hardware. It is easy to see the clear parallels between even such geographically distant jurisdictions.

Innovation driven by software

It is important to understand the gravity of these developments, since a significant majority of innovation is now driven by patenting software, as for example within AI, where specific self-learning algorithms represent it. Here, the technical step of the invention is also mainly fulfilled by associating the AI algorithm with actual technology, such as, this being a specific example provided by the EPO’s Enlarged Board of Appeal, a self-learning AI that monitors a pacemaker for the purpose of identifying irregular heartbeats. It’s interesting to see how many different technologies are now covered by AI-focused patenting, and in how many countries these patents are being granted.

Main publishing jurisdictions of patents mentioning AI, and secondary technologies they focus on

In summary, while the are significant hurdles and differences between different jurisdictions when it comes to applying to patent software, the relevant authorities have made it possible to provide sufficient evidence to connect it to a relevant technology and/or product, thereby still making it viable. Considering the massive impact software has generated within human innovation itself, it makes complete sense that its importance has not only already manifested itself within patenting, but will also undoubtedly continue to grow in the future.

I hope you enjoyed the software-focused article! Should you have any questions about the subject matter, or the related analytics in this article, all created by the Orbit Intelligence system of Questel SAS, don’t hesitate to reach out to us at help@questel.com. Some of the descriptions under the article’s analytics, have an automatic link attached to them, with additional data about how they were created and related support.