The European Patent Office (EPO) has just published the 2018 edition of its Guidelines for Examination in the European Patent Office. The date of entry into force of this latest version is 1 November 2018. This is in line with the EPO’s practice of updating these Guidelines on a yearly basis.

The Guidelines are made up of 8 parts, relating to:
– formalities examination;
– search;
– procedural aspects of substantive examination;
– opposition, limitation and revocation procedures;
– general procedural matters;
– requirements of European patent applications (e.g. sufficiency of disclosure, clarity, unity of invention, priority);
– patentability (e.g. patent eligibility, prior art definition, novelty, inventive step);
– amendments and corrections.

For the 2018 edition, changes have been introduced into all 8 parts of the Guidelines. In particular, various changes have been introduced to take into consideration recent decisions of the Enlarged Board of Appeal, i.e. the EPO’s highest instance, and of the Boards of Appeal. The new version is also the result of suggestions and comments received from users of the EPO system.

For example, the part relating to formalities examination now includes a mention of payment by credit card as a new payment form, and also other changes reflecting amendments to the payments via EPO deposit accounts. That same part now also includes a mention of the validation agreements with Tunisia and Cambodia dating back to 2017 and 2018 respectively.

Similarly, in the part relating to the requirements of European patent applications, the chapters addressing clarity of the claims and unity of invention have undergone significant revisions.

Another notable example is the part relating to patentability, where the chapter on patent eligibility has been revised and new subchapters relating to, inter alia, artificial intelligence (AI) and machine learning have been introduced. On this topic, the 2018 Guidelines state that AI and machine learning “are based on computational models and algorithms for classification, clustering, regression and dimensionality reduction, … Such computational models and algorithms are per se of an abstract mathematical nature, irrespective of whether they can be “trained” based on training data.” Thus, as a rule the EPO will carefully consider whether these models and algorithms are, albeit computerized, purely mathematical methods which, as such, are not patent-eligible under the European Patent Convention.

Nevertheless, the Guidelines also acknowledge that AI and machine learning “find applications in various fields of technology”. By way of example, the Guidelines cite using neural networks in heart-monitoring apparatuses for identifying irregular heartbeats and classifying digital images, videos, audio or speech signals based on low-level features as typical technical applications of classification algorithms which may thus be patent-eligible precisely thanks to their technical application. The Guidelines further explain that “[w]here a classification method serves a technical purpose, … generating the training set and training the classifier may also contribute to the technical character of the invention if they support achieving that technical purpose.

In view of the above changes and updates, which also reflect technological advancements, the 2018 edition of the Guidelines will surely represent an essential tool for all users of the EPO system.


pdf download PDF »

ISSN 2531-4483