Share this article:

Drafting Materials Related Applications Involving AI: Considerations for Success at the EPO

By Joanne Addison, Partner

The use of artificial intelligence (AI) in materials science is rapidly growing. AI can be useful across a range of technical areas in the field, such as: screening databases of known materials; materials modelling; materials design; and the prediction of properties of materials.

Patentability of AI inventions in materials fields

When considering the patentability of AI inventions, we need to assess whether the EPO will consider the claimed subject-matter to be technical and therefore not excluded from patentability. The EPO considers this requirement to be met if the AI invention is either: adapted to a specific technical implementation (in the sense that the design of the AI model is motivated by technical considerations of the internal functioning of the computer system on which it is run); or applied to a field of technology which the EPO has determined as being technical. For materials inventions involving AI, it is likely to be relatively straightforward to write claims that are considered to define technical subject-matter based on application to a specific field of technology. However, before we draft materials related patent applications which use AI to some extent, an important point to consider is the role AI plays in the invention.

What role does AI play in the invention?

Before we start draft drafting, we have to ask some key questions about the nature of the invention.  For example, is the AI aspect of the invention incidental to the invention or is the invention only possible because of the advent of AI? Or does the invention actually represent a contribution to the field of AI itself rather than just to a specific field within materials science? The answers to these questions help us distinguish between “Applied-AI inventions” and “Core-AI inventions”.

Applied-AI inventions

I’m using the term “Applied-AI inventions” here to refer to inventions which use a known AI algorithm to solve a problem in a particular area of materials science, i.e., where the invention lies in the manner in which a known AI algorithm is used. These types of inventions are considered technical (and so not excluded from patentability) by the EPO due to being “adapted to a field of technology”.

We can split Applied-AI inventions into two main categories: inventions where AI is incidental to the invention; and inventions which are only possible because of the advent of AI. The category of AI invention impacts the content that may be required in the claims and description of a patent application.

An example of a materials invention where AI is incidental to the invention and merely one way that the invention might be realised could be an invention involving a step of simulating the behaviour of materials under different conditions, in which the behaviour is simulated using a machine learning model (but might also be performed using other methods). For inventions of this type, AI is an implementation detail, but not the main invention. Therefore, when drafting a patent application for this type of invention, AI embodiments would be unlikely to be useful in the independent claims, possibly even the dependent claims. It may, nevertheless, be helpful to include details of the AI implementation in the description in order to provide details of how to work the invention.

AI has, however, also opened up many possibilities in materials fields. The second category of Applied-AI inventions are those which, whilst not representing improvements to fundamental AI algorithms, are only possible because of the advent of AI. In materials science, an example of such an invention could be using AI to predict combinations of elements which will form structures having desired properties. For example, it may be that producing a new type of alloy from many different elements to provide a material having a particular property would not have been contemplated, or indeed possible, before the advent of AI and so such a method could be considered inventive. Another example of this second type of Applied-AI/materials invention could be one that involves identifying a property (e.g. crystal structure) of a material from measurements of other seemingly unrelated properties. For drafting purposes, as the technical effect of these types of invention can’t reasonably be obtained without AI, the AI will likely feature in the independent claims of a patent application. We would also expect the description to include further details regarding the AI embodiments for the purposes of sufficiency and inventive step (see below for further details).

It is worth noting that if a known AI algorithm is simply used to improve or automate a known process, for example, solving crystal structures from diffraction patterns using known machine learning methods, it may be difficult to demonstrate inventiveness in Europe. Before proceeding with drafting an application directed to such subject-matter, it should be considered whether the use of the known AI algorithm produces an unexpected technical effect, or if there are any other merits which would demonstrate the presence of an inventive step.


Here the term “Core-AI” is used to refer to inventions which represent a contribution to the field of AI itself. For example, better models, improved pre-processing of data, improved methods of training. These inventions can generally be applied to a wide range of problems across a wide range of fields.

It may be initially assumed that inventions in materials fields that use AI will usually fall into the category of “Applied-AI” and are unlikely to ever make a “Core-AI” invention. However, we would caution against this assumption, as many inventions in fundamental aspects of AI arise from real-world problems that inventors encounter when trying to apply known techniques to their particular data.

For example, an inventor may have found a way of reformatting their data after finding the data set was too small for the model to successfully operate. If the same technique can be used on other small datasets, then this is a Core-AI invention.

Another practical example is advancements in federated learning which arise due to organisations needing to pool their datasets in order to train a model, without wanting to actually share the underlying data directly.

While inventions relating to Core-AI are generally more likely to encounter patentability issues at the EPO, in particular mathematical method objections, when these inventions are developed based on problems encountered in materials fields it will be possible to clearly describe how the invention is adapted to a field of technology, by providing use-cases. In Europe, even Core-AI inventions may need to be limited to a use-case to ensure the claims are considered technical (and so not excluded from patentability) due to being “adapted to a field of technology”.

In the first instance, inventions relating to Core-AI may be defined usefully by independent claims directed to the mathematical method itself, irrespective of any field restrictions. This ensures that the application offers the applicant the widest number of options in prosecution, particularly if the application will also be prosecuted in other jurisdictions having different patent eligibility requirements to the EPO.

However, with the EPO in mind, the dependent claims should contain specific use-cases of different, and preferably graded, scopes. Importantly, a conversation should be had about the realistic scope that may be obtained for Core-AI inventions in Europe and the commercial usefulness of such claims.

With this in mind, use-cases might be chosen that detail how the Core-AI invention could be applied to the applicant’s most commercially important AI models and products. They should therefore be chosen in a strategic manner, as opposed to merely relying on the use-cases provided by the inventors themselves. Building up a claim set in this way provides the best opportunity to obtain granted European patents that are commercially relevant for Core-AI applications.

Sufficiency and Inventive Step Considerations

In materials fields, providing experimental data to ensure that patent applications meet the requirements of sufficiency at the EPO and demonstrate an inventive step is usually a fairly major consideration when drafting. However, it should be noted that for inventions in materials fields in which AI is a key aspect, the provision of examples is also important in terms of sufficiency and inventive step of the AI aspects.

Researchers in the field of AI understand that the design of a training data set can be critical to success of an algorithm, as well as the possible effects of model assumptions and design. Therefore, if AI embodiments are to be included in the claims, the patent application should provide information in relation to the training data set (e.g., size, how outliers are handled, selection), the model used to derive the AI (including the type of model, e.g. neural network, genetic algorithm, a decision tree, etc. and how the model is structured), as well as any assumptions made by the model.

Furthermore, if, as suggested above, it is necessary for a “use-case(s)” to be defined in the claims in order to meet the subject-matter eligibility requirements in Europe, this will effectively result in a technical effect being defined in the claims. When a technical effect is defined in a claim, in order for the requirement of sufficiency of disclosure to be satisfied in Europe, it is necessary for the technical effect to be made plausible to the skilled person at the filing date, based on the information provided in the application as filed and their common general knowledge. Therefore, even for Core AI inventions, it is likely to be necessary to provide enough evidence in the application to make the “use-cases” described in the application plausible (e.g., evidence that this technical effect is achieved by the claimed invention). Without such evidence, the application could be considered to be incurably insufficient, or it may be necessary to limit the claims to a more specific “use-case” described in the application (e.g. a more narrowly defined technical effect) that is considered to be made plausible by the evidence provided in the application or the skilled person’s common general knowledge at the filing date.


Materials inventions involving AI can be patentable in Europe unless the invention is purely directed to using a known AI algorithm to straightforwardly improve or automate a known process. However, it is important to consider the role AI plays in the invention when drafting the application in order to formulate a suitable breadth of claim and provide suitable evidence for the purposes of sufficiency and inventive step in Europe.

Contact Jo Addison, or any of our AI team members, at Haseltine Lake Kempner if you have any questions or would like to discuss these issues further.

This is for general information only and does not constitute legal advice. Should you require advice on this or any other topic then please contact or your usual Haseltine Lake Kempner advisor.

HLK bubble graphic HLK bubble graphic

Stay connected with HLK

Keep up-to-date with the latest IP insights and updates as well as upcoming webinars and seminars via HLK’s social media.