Once only a figment of the imagination in literature such as Mary Shelley’s Frankenstein, artificial intelligence, commonly known as AI, is now realised and thriving in many technical areas such as medicine, finance, transport and even advertising. Considered by many as one of the most important technologies of the future, global investment in AI is growing at an ever increasing rate, with AI revolutionising the way we all live, work, and communicate.
Is it possible to get patent protection for AI technology?
The short answer is yes.
In fact, the EPO’s latest edition of the Guidelines for Examination explicitly states that artificial intelligence can be patentable at the EPO. The Guidelines also provide more specific guidance on the examination of AI applications; specifically that the EPO will examine AI applications under the EPO’s existing approach to computer-implemented inventions (CIIs). This existing approach is stable and predictable, backed by a solid body of case law. Indeed, we believe that one purpose of the updates to the Guidelines for Examination is to actively advertise to applicants that the EPO is “open for business” for AI patents.
The patentability of CIIs at the EPO has been governed since 2002 by the principle that there must be a non-obvious technical solution to a technical problem (T0641/00, COMVIK). Only the technical features in the claim can contribute to inventive step; the non-technical features (such as purely abstract mathematical or business methods) cannot.
Of course, the foundation of much AI technology is likely to be based on computational models and algorithms which are per se of an abstract mathematical nature. The new edition of the Guidelines clarifies how inventions comprising mathematical methods (such as AI inventions) can be patentable.
How can I get patent protection for AI technology?
Mathematical methods may contribute to the technical character of an invention (i.e. contribute to producing a technical effect that serves a technical purpose) by:
- its application to a field of technology; and/or
- being adapted to a specific technical implementation.
This means that AI as such is not considered technical. For example, classifying abstract data records or providing computer or memory efficiency inherent in an AI algorithm is not enough to overcome the hurdle to patentability.
But, according to point 1 above, if the AI technology is applied to a specific technical field, there is a technical purpose which can result in patentability.
Accordingly, examples of AI that are patentable include use of a neural network in a heart-monitoring apparatus for the purpose of detecting irregular heartbeats, or classification of digital images, videos, audio or speech signals based on low level features. Other examples of mathematical methods that are patentable include controlling a specific technical system or process (e.g. an X-ray apparatus, a steel cooling process), a system encrypting/decrypting or signing electronic communications, or a system for encoding data for reliable and/or efficient transmission or storage.
Therefore, in a patent application, claims directed to an AI invention and restricted to their specific technical purpose should be patentable. A generic purpose such as “controlling a technical system” is not sufficient to confer a technical character to the mathematical method. The purpose must be specific to a technical field.
According to point 2 above, the design of the AI technology must be motivated by technical considerations of the internal functioning of the computer in order to provide for patentability. Therefore, AI technology which is specifically adapted or designed to run on particular hardware should in principle also be patentable.
Here to help
Patent protection at the EPO for AI technology is possible. And we are here to help.
We are seeing a substantial increase in the number of applications we are filing involving AI technology, and we have a wealth of experience in the drafting and prosecution of AI inventions. It is clear that the growth in investment and development in AI technologies is on course to boom in the near future. Those who seek patent protection for their AI inventions now, will be ready.
About the author
Lucy Coe is a member of our electronics, computing & physics and engineering patent teams. She has an MPhys (Hons) degree in Physics from the University of Manchester. Her Masters project focused on assessing the feasibility of building an interactive table-top particle accelerator to demonstrate strong focusing and included a complete design and proposed method of manufacture.