Our services are centred around intellectual property that can be registered. We protect innovation, design, and branding across all sectors of industry, and at all stages in the supply chain.

For each IP right we offer services covering strategic advice, pre-registration searches, registrations and renewals, oppositions and dispute resolution. We handle work throughout the world, working with local colleagues in over 100 countries.


Our attorneys specialise in one or more sectors of industry, which enables them to provide quality advice with a commercial focus.

Our patent specialists have detailed understanding of the background technology, which ensures that your patent applications are prepared with the correct scope, reducing the likelihood of challenges from third parties and objections from the patent office.

They also advise whether other forms of protection would be more appropriate. Our brand specialists work with brand managers for leading brands and their advice is commercially focussed making sure that you get the best value from your budget.

Patenting artificial intelligence and neural networks in Germany


Patents relating to artificial intelligence or machine learning are currently in the spotlight and have a strong increase in the number of thereon filed patent applications, even though there are often high hurdles to overcome before granting them. Reason enough to take a closer look at a current case of the German Federal Patent Court (Bundespatentgericht – BpatG) regarding an appeal in patent examination proceedings of a patent application dealing with trained artificial neural networks.

The German Federal Patent Court approached in the decision No. 19 W (pat) 7/22 dated 1 June 2022 the question of patentability of inventions dealing with artificial intelligence in accordance with its established legal practice regarding the patentability of computer-implemented inventions.

The invention as disclosed by the patent application allows a driver assistance system to visually recognize a person’s wish to be transported by the passenger transport vehicle and the disclosed invention further allows to automatically steer the vehicle to the person indicating the wish by a gesture using a trained artificial neural network. Thus, artificial intelligence was used to improve the human-machine interaction and to enable advanced, vision-based gesture recognition.

At the end of the proceedings before the examination departments of the first instance, the patent application was rejected by the German Patent Office for lack of inventive step.

At second instance before the appeal court, the applicant argued with regard to inventive step that machine learning cannot only be realized by algorithms with the architecture of an artificial neural network as claimed, but also by statistical classifications methods using classifier networks, by other algorithms, such as decision trees, artificial decision trees, the so-called k-means algorithm, or support vector machines which can be used as pattern recognition and classification of objects for the mentioned application. The applicant stated that the selection would be connected to the particular technical effect of an improved solution and the skilled person would not have been prompted to the specific selection due to the absence of any hints.

The German Federal Patent Court dismissed the appeal and denied the inventive step of the claimed subject matter, as the use of a certain type of a neural network in terms of an artificial neural network was considered to be common general knowledge and selecting one specific type of a neural network out of a variety of suitable neural network algorithms was considered to be obvious by the court.

Interestingly, the court did not question that the features defining the human-machine interaction or the artificial neural network are technical. Probably, the question of technicality was not raised, because the underlying problem was formulated as a classical sensor problem how to recognize the typical transport wish gestures of persons.

The court further did not release the decision for appeal to the Federal Court of Justice (Bundesgerichtshof – BGH). This can also be seen as an indication that the court had no open questions regarding the technical character of the invention or even of the single features defining the artificial neural networks.

As a conclusion for drafting, it remains to be said that according to German patent prosecution practice for a patent to be granted, it would have been probably necessary to better link any specific implementation or adaption of the artificial neural network to the present object of an automated and vision-based recognition of a transport request. As a result, off-the-shelf solutions that merely consist of the first-time application of artificial intelligence in a certain technical field for a specific technical task are not sufficient to overcome the hurdle of inventive step.

On a positive note, the court’s decision reads, at least implicitly, like an affirmation of the technical character of artificial intelligence systems in general, as a deeper discussion of any software patent eligibility is lacking in the reasoning and all features of the independent claims were substantively examined by the court for inventive step and thus considered technical.

The decision shows that the prosecution of patents on artificial intelligence via the national route in Germany can be advantageous insofar as now apparently only decreased barriers are to be expected with regard to the classification of features into technical or non-technical features or with regard to the judgment of the technical character of an invention per se.

This article is for general information only. Its content is not a statement of the law on any subject and does not constitute advice. Please contact Reddie & Grose LLP for advice before taking any action in reliance on it.

Saved Staff
Staff member

Remove all

Call +44 (0)20 7242 0901
Call +44 (0)1223 360 350
Call +49 (0) 89 206054 267
Call +(00) 31 70 800 2162
This field is for validation purposes and should be left unchanged.