Patentability of AI innovations

This is an Insight article, written by a selected partner as part of WTR's co-published content. Read more on Insight

Examination standards – AI patent application strategies

Thanks to the rapid development of AI technology and continuous reflection on its commercial values, patent applications related to AI technology have become a hot topic. The number of AI-related applications is rising, while the scope of the application fields is expanding.

This chapter aims to provide some patent application strategies in the AI field based on the latest examination standards in China, summarising the similarities and differences between those standards in China, Japan, Korea, the United States and Europe.

Laws and coping strategies

In China the primary examination focus of a patent application relating to AI looks at whether the application is for an eligible object to be protected by a patent and whether its inventiveness is as provided in Article 22(3) of the Patent Law.

A patent application in the AI field may be drafted as a product claim or a method claim. A product claim may be drafted as an eligible subject, such as a system, device or storage medium.

Figure 1: General examination process of a patent application in AI field

Table 1: Forms of drafting claims

	Eligible subject	Example of drafting
Method claim	Method	A speech recognition method based on a deep neural network, comprising step 1 and step 2.
Product claim	Apparatus (device), system	A program module – a speech recognition apparatus based on a deep neural network, comprising module 1 configured to … (corresponding to step 1); and module 2 configured to … (corresponding to step 2). Memory plus a processor – a speech recognition apparatus based on a deep neural network, comprising a memory configured to store a computer program; and a processor configured to carry out the following steps when the computer program is executed: step 1; and step 2.
Product claim	Storage medium	A computer readable medium, including a computer readable program code, which will cause a computer to carry out the following steps: step 1; and step 2.

Examination standards and coping strategies

Latest examination standards on eligible object issues

Article 25(1), Item 2 of the Patent Law provides that a patent right will not be granted for rules and methods for mental activities.

The newly amended Guidelines for Examination provide that if a claim contains a technical feature in addition to an algorithm feature or a commercial rule and a method feature, the claim as a whole is not a rule and method of an intellectual activity. The possibility that it will be granted a patent right should not be excluded in accordance with Article 25(1), Item 2. Moreover, it is provided in Rule 22(2) of the Implementing Regulations of the Chinese Patent Law that “invention as mentioned in the Patent Law means any new technical solution relating to a product, a process or an improvement thereof”.

According to the guidelines, if, in a claim, the steps involved in an algorithm reflect that they are closely related to the technical problem to be solved (eg, data processed by the algorithm has definite technical meanings in the technical field), that execution of the algorithm can directly reflect a process of solving a technical problem by using natural laws, and produce a technical effect, then, in general, the solution defined in the claim belongs to the technical solution provided in Article 2(2) of the Patent Law.

Application strategy for eligible object issues

AI patent applications can be divided into two types according to their application scope:

basic type patent applications; and
applied type patent applications.

The ‘basic type’ patent application refers to the fact that an algorithm involved in the application may be widely used in multiple particular fields. The ‘applied type’ refers to the fact that an algorithm involved in the patent application is mainly combined with a particular field and is an application in this field.

Taking two aspects into account (ie, patent protection scope and conformity to examination requirements), there are a number of possible ways for drafting the two types of patent application.

Due to the development of the Internet and big data technology, AI technology is also being used increasingly in commercial and financial fields. In making an application for this type of patent, attention should be paid to combining a business rule, an algorithm feature and a technical feature in the description.

Moreover, based on a stage of technological improvement, an AI patent application may be divided into two stages:

the training (or learning) stage; and
the application stage.

Table 2: The different ways to draft a patent application

		Basic type	Applied type
Claim	Protected subject	Field is more generic	Field is more specific
Claim	Technical solution	Processed object (input data) – the input data has specific meanings and may be technical or non-technical data. Means of processing – combining an algorithm feature and a technical feature. Result of processing (output data) – a result of processing having a technical meaning.	Processed object (input data) – the input data is specific data in a specific field. Means of processing – combining an algorithm feature and a technical feature. Result of processing (output data) – a result of processing having a technical meaning.
Description	Technical problem and technical effect	Attempt to summarise based on improvement on performance/ precision/efficiency.	A specific problem in a specific field.
Description	Embodiment	Drafting multiple typical application examples regarding different specific application fields, and describing specific technical effects in various specific scenarios in embodiments.	Application examples in a specific field.

Examination standard and coping strategy regarding inventiveness

Latest examination standards regarding inventiveness

According to the recently amended Guidelines for Examination, when examination of inventiveness is conducted for a patent application for an invention containing a technical and algorithm feature, or a business rule and a method feature, the algorithm feature or the business rule and method feature will be taken into account together with the technical feature as a whole, when they functionally and mutually support the technical feature and have an interactive relationship between themselves and the technical feature.

Application strategy for examination on inventiveness

Based on the above examination standards, when an application for an AI patent is drafted, attention should be paid to combining an algorithm feature and a technical feature when describing the technical solution. Moreover, in describing a technical problem and a technical effect, emphasis should be placed on the fact that the algorithm feature and the technical feature are specifically combined and jointly solve the technical problem and produce a corresponding technical effect.

Further, for some AI patent applications not involved in the improvement of a basic algorithm, any improvement is relative to the existing technologies in the application of an algorithm (eg, a neural network) to a specific field, while the neural network itself is not changed much. For this type of patent application, inventiveness may be considered based mainly on:

whether the technical fields are similar; and
the difficulty of applying the neural network to the technical field of the present application and whether a technical effect different from that in the original technical field is produced.

Table 3: Eligible subjects in two stages

Improvement	Eligible subject	Examples
Training (learning) stage	A training apparatus/method for a model/neural network	A training apparatus/method for an image recognition model
Application stage	An apparatus/method for (a function)	An image recognition apparatus/method
Training stage and application stage	Both of the above	Both of the above

Comparison of examination standards

Eligible object protected by a patent

For a comparison of examination standards of an eligible object protected by a patent in China, Japan, Korea, the United States and Europe, see Table 4.

Table 4: Examination of an eligible object protected by a patent in China, Japan, Korea, United States and Europe

	Subjects that can be protected	Requirements on technical solutions of claims
China	An apparatus/system/method. A storage medium.	Must be applied to a specific field, and a combination of an algorithm feature and a technical feature should be reflected.
Japan	An apparatus/system/method. A computer program (drafted as “a computer program enabling a computer to execute a function of…”). A storage medium. A model completed by training.	An implementation carried out specifically by using a hardware resource that is able to reflect that information processing is based on a computer and that it is an advanced creation of a technical idea using the natural law.
Korea	An apparatus/system/method. A computer program recorded on a storage medium. A storage medium.	An information processing procedure of AI may be specifically carried out by hardware. Further, for an application for a patent for applying an AI technology for disease diagnosis, examination standards are appropriately relaxed.
United States	An apparatus/system/method. A computer program product (including or included in a computer readable medium + non-transitory). A storage medium.	A judgment method with reference to computerrelated applications for patents: judicial exception (abstract concept, natural law or natural phenomenon) is not cited; judicial exception is cited, but is integrated into practical use; or judicial exception is cited and the judicial exception is not integrated into practical use, but any component or a combination of components is sufficient to ensure that the claim is significantly more than the judicial exception itself.
Europe	An apparatus/system/method. A computer program. A computer program product. A storage medium.	Similar to that in China, applied to a specific field, and an algorithm feature and a technical feature are combined.

Inventiveness

For a comparison of examination standards of inventiveness in China, Japan, Korea, the United States and Europe, see Table 5.

Table 5: Examination of inventiveness in China, Japan, Korea, United States and Europe

	General examination of inventiveness	Specialty of AI field
China	Based on a three-step method including: determining closest prior art; determining a distinguishing feature and a technical problem to be actually solved by the present application; and judging whether a claimed invention is obvious to a person skilled in the art.	An algorithm feature and a technical feature which are functionally and mutually supporting and have an interactive relationship to be taken into account as a whole.
Japan	Compare the present application with the closest reference to find out an identical technical feature and a distinguishing technical feature. Perform inference on the distinguishing technical feature.	Examples of possessing no inventiveness: systematising a human activity by using AI technology; an improved method in which an output result is predicted based only on input data; and amendments to training data used for machine learning are merely a combination of known data and no outstanding effect is produced. Examples of possessing inventiveness: pre-processing of training data used for machine learning; and bringing outstanding effect by training data added for machine learning.
Korea	Judge whether a person skilled in the art is able to easily obtain the invention concerned in the present application according to a reference prior to the present application and common general knowledge.	A judgment method with reference to computerrelated applications for a patent – considering technical difficulties of applications of a related technique in different fields, whether a solved problem is a common technical problem in the computer field and whether a produced effect is a common technical effect in the computer field.
United States	Determine a difference between a claimed invention and the prior art, analyse a level of a person of ordinary skill in the relevant field, and judge whether the difference is obvious with reference to objective evidence.	A judgment method with reference to computerrelated applications for patent – proposing an abstract concept issue together with an inventiveness issue, and after the problem of abstract concept is overcome by making amendments (eg, combining with a specific field), the problem of inventiveness may possibly be overcome at the same time.
Europe	Similar to the three-step method in China, adopt a problem-solution method, including three steps: determining closest prior art; determining an objective technical problem to be solved when a distinguishing feature brings a technical effect; and starting from the closest prior art and the objective technical problem, consider whether an invention to be protected is obvious to a person skilled in the relevant technology field.	Similar to that in China, consider whether an algorithm feature makes a technical contribution.

Comment

Patent applications in the AI field belong to patent applications involving computer programs, which must therefore meet the universal requirements for patent applications involving computer programs. Due to the specialised nature of AI technology, for patent applications in the AI field, the China National IP Administration has formulated new special examination regulations. Drafting patent applications and the responses to examination opinions based on the latest examination standards are beneficial to applicants in obtaining patent rights of relevant technologies in China.

In addition, understanding examination standards for patent applications in the AI field in major patent countries and regions in the world (eg, China, Japan, Korea, the United States and Europe) is advantageous to global application strategy formulation and reasonable patent layout of applicants.