The Chartered Institute of Patent Attorneys (CIPA) in the UK has called for clarity over the patenting of innovations created by artificial intelligence. The issue has drawn a lot of attention lately. Last January, the EPO stated a machine cannot be an inventor, refusing patent applications EP 18 275 163 and EP 18 275 174 in which a machine called DABUS was named as the inventor.

In a position paper published this week, CIPA wrote: “Until very recently, human intellectual and practical endeavour has been responsible for creating (…) inventions (…). However, as the cognitive capabilities and power of artificial intelligence (AI) systems improve, they are already participating in advances across a wide range of technical fields, including medical research, such as drug discovery, and autonomous vehicles. An invention may be created using an AI system that will challenge this human-centric view of inventorship.”

The paper refers to consultations and studies by the USPTO, the EPO and the UK Intellectual Property Office (IPO). In the decisions mentioned above, the EPO “considered that the interpretation of the legal framework of the European patent system leads to the conclusion that the inventor designated in a European patent must be a natural person. The Office further noted that the understanding of the term inventor as referring to a natural person appears to be an internationally applicable standard, and that various national courts have issued decisions to this effect.”

According to the CIPA paper, there is “a tension between a desire by some applicants to obtain patent protection for an invention apparently created solely by an AI system, and the current legal position as expressed by various patent offices.”

“Many in CIPA think patent rights should be available for inventions which represent new, non-obvious technical developments, regardless of how they were created (with or without an AI system).

Others in CIPA prefer to limit patent protection to inventions having a human contribution – in effect, retaining current inventorship requirements, but accepting that an invention created using AI is patentable as long as there is a genuine human contribution.”

CIPA calls for an investigation of the issue, “including discussions with stakeholders such as industry, policy-makers and legislators”, as “the ongoing uncertainty might cast doubt on the validity of granted patents for inventions created using an AI system; it might also potentially impact AI supported research and development in the UK and elsewhere.”


To make sure you do not miss out on regular updates from the Kluwer Patent Blog, please subscribe here.

Kluwer IP Law

The 2022 Future Ready Lawyer survey showed that 79% of lawyers think that the importance of legal technology will increase for next year. With Kluwer IP Law you can navigate the increasingly global practice of IP law with specialized, local and cross-border information and tools from every preferred location. Are you, as an IP professional, ready for the future?

Learn how Kluwer IP Law can support you.

Kluwer IP Law
This page as PDF


  1. AI generated patents, of the patent system will collapse under the number of patent applications.

    1. Until someone invents the AI that does prior art searches and inventiveness/obviousness examination.

      Will be fun to see the amount of inter-party discussions resulting from those grants and refusals.

  2. The problem of AI inventions is not valid only for the IPO, but also for the EPO. That CIPA wants a clarification is nothing surprising. But is such a clarification really needed?

    Two aspects have to be considered, and they are quite distinct.

    The first one is about ownership of an invention allegedly resulting from AI. Can the ownership of an invention be left to a machine? I have strong doubts about this. May be, the ownership of the AI invention could be given to the owner of the neural network used to “create” the invention.

    In order to circumvent the problem of inventorship, the name of a physical person could be given as inventor when filing an application for patent, and hence not saying who the true inventor is. As it is only the result which is patented, and provided the application gives the person skilled in the art the possibility to carry out the invention, there is no apparent need to indicate that the patent is the result of an AI machine.

    The inventor has a moral right to be named, cf. Art 81EPC, but according to R 19(2)EPC, the EPO “shall not verify the accuracy of the designation of the inventor”. Does a machine have a moral right? Strong doubts are permitted. One could therefore even wonder whether the refusal of the application

    The second aspect which is quite distinct from the first one, is the patentability of AI when it helps to resolve a technical problem.

    That AI can help when it comes to crunching masses of data is not at stake here. One example could be mass screening of X-ray pictures in order to detect malignancies. Object, not to say target, recognition could be another one. Inventions based on AI can be dealt with like any other CII, even if some people disagree with this.

    Going back to the famous Vicom case (T 208/84), it could be looked at as an AI invention, as the repetition of the convolution of a specific kernel with image data allowed to increase contrast in the original picture. AI is no more than the repetition of a specific action with a set of data in order to obtain a given result.

    However, in order to obtain a patent, it will at least be necessary to indicate the input data that allow the network to learn, as well as the correlation rules. Without these two elements there is no sufficient description. If the learning data or the correlation rules change, the output result changes.

    It is hard to see how all this can be described as intelligent. And that also applies to invention allegedly obtained by AI.

    There is thus nothing intelligent in these kind of machines, they only do what they are told to do. They are only doing what they are told and if some self-perturbation of connection weights between neurons, like alleged in DABUS, should all bring the desired result, this needs a bit more explanation.

    As far as the EP 18275163 (fractal food container) and EP 18275174 (fractal light signals) patents are concerned, the following can be said:

    For EP 18275163, it is known to connect cans by their external profile. The only difference is that in the case of the application, the connecting surface is a fractal surface. It remains to be seen whether this could be inventive.

    EP 18275174 is at the verge of a lack of sufficiency, in that the fractal light emission relies on a theory devised by the applicant himself, theory which seems rather obscure.

    A general comment:

    AI seems to be very fashionable. I would even speak from a hype, but this hype will like other ones (biotech, big data) die as all its predecessors did. That there may be practical applications is quite certain. But to translate into patents, more than some vague indications seem necessary.

    AI is just another expression for “Neural Networks”, and it is the vast increase in the capacity of data processing that makes them now more usable.

    AI is also a mean to get plenty of public money, so why not jump on the bandwagon?

    It is also a wonderful playground for academics which ravel in such theoretical constructs. It reminds me of lots of papers, if not books, written by lawyers in the early days of the EPC about inventive step. How much did all those essays have as practical utility when it came to assess inventive step in real life? I spare you a reply. In the meantime, the problem-solution-approach has taken over, at least at the EPO.

    Some people, like the former president of the EPO go even as far as to claim that in a few years, drafting patent applications will be taken over by AI machines, and the same applies to the job of an examiner, be it search or examination, and even opposition. May be outside the solar system, but I do not think that will happen in his lifetime or even that of his grandchildren should he have any.

    Techrights and zoobab: FINGERS OFF!!

    1. Numerous comments have been submitted in response to the USPTO’s consultation by various stakeholders, including the EPO, companies, individuals, industry organisations, IP professionnal associations. A summary review of these valuable comments would be helpful to shed light on the issue. Special attention should be given to a study commissioned by the EPO, authored by Dr Noam Shemtov (” A study on inventorship in inventions involving AI activity”), mentioned in the EPO’s comments. The analysis and conclusions of this study are quite clear.
      However, there may be reasons within some UK circles for signalling flexibility as to the possibility of avoiding the designation of natural persons as inventors even though (or just because) there is worldwide unanimity on this issue.

  3. My view with regard to AI is that if there is nothing “clever” in designing or programming the system, then no invention can ever result as any output is an automatic result of a non-inventive activity. If, on the other hand, the designing or programming is inventive, then the designer or programmer is the inventor who should be identified.

Comments are closed.