China’s New Guidelines – Good News for Software Patents

Did you know?

The Chinese National Intellectual Property Administration (CNIPA) recently revised the Examination Guidelines for software patents including artificial intelligence (AI), big data and blockchain.

Why does this matter to you?

These new Examination Guidelines provide much needed clarity as to the boundaries of patentable subject matter in China.

How are software patents treated in other jurisdictions?

The patent laws which we have today were originally designed to protect the products of the 19th century industrial revolution: mechanical devices and chemical compositions. As such, software never sat easily within the patent system and was initially barred from patentability in many jurisdictions, including China. This became untenable as software became ubiquitous and entered every aspect of the economy, resulting in a complex body of US and European case law providing numerous exceptions, which allow software to be patented in certain circumstances. 

The European Patent Office has stated that software can be patented if it solves a “technical problem”, while the US allows software to be patented if the claims recite more than an “abstract idea”. However, the terms “technical” and “abstract” are not well-defined and the European and US Patent Offices often interpret these restrictions broadly, which results in a lack of certainty and can make it difficult to obtain software patents.

What is the approach in China?

The Chinese Examination Guidelines previously required a software patent to use “technical means” to solve a “technical problem” and provide a “technical result” (sometimes referred to as a “triple technical requirement”). The new Guidelines provide more substantive guidance and recast the criteria as a two-part test.

Firstly, the Examiner determines whether the claim recites rules and methods for mental activities only. If so then, the claim is unpatentable under Article 25 of the Patent Law. However, if the claim includes rules and methods for mental activity in combination with technical features, then the first hurdle is passed.

The second hurdle is compliance with Article 2 of the Patent Law, which requires that the claim defines a “technical solution”. A claim is considered to define a technical solution if it recites technical means which solve a technical problem to achieve a technical effect. While this is similar to the previous Guidelines, the new Guidelines stress that when making this assessment, all features of the claim have to be considered as a whole.

Therefore the Examiner cannot disregard non-technical features of the claim, such as features relating to an algorithm or business method, even though they are not technical by themselves, as they may interact with the technical features of the claim to solve a technical problem. This is a welcome development as the inventive part of a claim often resides in the so-called “non-technical features”, especially if the invention is implemented on standard computer hardware.

What are some practical examples?

The new Guidelines provide several illustrative examples. An example of a claim which fails the first hurdle is a mathematical method which includes detailed processing steps, but is not limited to any specific application area. Therefore, to pass the first hurdle, it is preferable to mention a particular application or type of technical data (e.g. image data). An example of a claim failing the second hurdle is a method of providing customer rebates for increased consumption, as this is considered to solve an economic, rather than technical problem. 

An improved design of convolutional neural network (CNN) providing better image recognition is given as an example of a claim passing the second hurdle. Other examples include a bike rental service which uses a server to provide information on rental bikes near a user’s smartphone; a blockchain system which uses a CA certificate and trust list for improved user security; and a logistics distribution method, which notifies customers within a certain delivery range by batch, rather than one by one. With the increasing importance of the Internet of Things, blockchain and e-commerce, the inclusion of these examples suggests that the CNIPA will be more willing to grant patents in these areas in the future.

Now that the technical character threshold has been lowered, software patents in China are rarely objected to because the claimed subject matter does not meet the statutory requirement.  However, from recent experience, the challenge has shifted to proving the presence of an inventive step.

If you have any questions, or require assistance patenting software, AI, block chain or big data inventions in China, please do not hesitate to contact us.