On February 3rd, the European Patent Office published a preview of the 2025 EPC Guidelines, which are scheduled to come into effect on April 1st.
The EPO is accepting comments on this preview version until April 7th via the following page, and the comments collected will be taken up at the SACEPO meeting on May 8th.
https://www.epo.org/en/law-practice/consultation/ongoing#GL2025
Below are some of the changes made to the 2025 edition of the EPC Guidelines.
Source:https://link.epo.org/web/legal/guidelines-epc/en-epc-guidelines-draft-2025.pdf
*The Japanese translation is for reference only.
General Part 5: On the responsibility of the parties to fulfil EPC requirements even when submissions are prepared by AI
The parties and their representatives are responsible for the content of their patent applications and submissions to the EPO and for complying with the requirements of the EPC regardless of whether a document has been prepared with the assistance of an artificial intelligence (AI) tool.
Parties and their representatives are responsible for the content of their patent applications and submissions to the EPO and their compliance with the requirements of the EPC, regardless of whether the documents were prepared with the assistance of artificial intelligence (AI) tools.
Part BX, 11.1: Availability of documents in search reports
Part C IV, 7.5: Access to documents not mentioned in the search report but newly mentioned in the examination
MyEPO Portfolio users receive all cited documents electronically in their Mailbox. Applicants not having opted for electronic notification via Mailbox receive paper copies of non-patent literature and translations of cited patent literature by post. Digital copies of cited patent literature documents are accessible in Espacenet (worldwide.espacenet.com/).
MyEPO Portfolio users will receive all cited documents electronically in their mailbox. Applicants who have not opted for electronic notification via their mailbox will receive paper copies of non-patent literature and translations of cited patent documents by post. Digital copies of cited patent documents can be accessed via Espacenet.
Part G II, 3.3.1: Clarification of the definitions of artificial intelligence and machine learning
*The underlined parts are additions made to the 2025 edition.
Artificial intelligence and machine learning are based on computational models and algorithms such as artificial neural networks, genetic algorithms, support vector machines, k-means, kernel regression and discriminant analysis. Such computational models and algorithms are per se of an abstract mathematical nature, irrespective of whether they can be “trained” using training data. However, their use does not by itself render inventions related to artificial intelligence or machine learning non- patentable, and the guidance provided in G-II, 3.3 generally applies. 52(2) or (3). In such cases, the computational models and themselves algorithms contribute to the technical character of the invention if they contribute to a technical solution to a technical problem, for example by being applied in a field of technology and/or by being adapted to a specific technical implementation.
Artificial intelligence and machine learning areArtificialThey are based on computational models and algorithms such as neural networks, genetic algorithms, support vector machines, k-means, kernel regression, discriminant analysis, etc. Such computational models and algorithms have abstract mathematical properties in themselves, whether or not they can be "trained" using training data.However, their use does not in itself render an invention related to artificial intelligence or machine learning unpatentable.The guidance provided in ,G-II,3.3 generally applies.This means that if a claim of an invention related to artificial intelligence or machine learning is directed to either a method or an apparatus involving the use of technical means (e.g. a computer), the subject matter as a whole has a technical character and is therefore not excluded from patentability under Article 52(2) or (3). In such cases, the computational models and algorithms themselves contribute to the technical character of the invention if, for example, they are applied in a technical field and/or adapted to a specific technical implementation, thereby contributing to a technical solution to a technical problem.
Part H VI, 2.2: Clarification of practices regarding admissibility in line with G XNUMX/XNUMX
For a correction of linguistic errors, errors of transcription and mistakes in any document filed with the EPO to be allowable, the following requirements must be met (G 1/12):
(i) The correction must introduce what was originally intended. The possibility of correction cannot be used to enable a person to give effect to a change of mind or development of plans. It is the party's actual rather than ostensible intention which must be considered.
(ii) Where the original intention is not immediately apparent, the requester bears the burden of proof, which must be a heavy one.
(iii) The error to be remedied may be an incorrect statement or an omission.
(iv) The request for correction must be filed without undue delay.
In order for the correction of linguistic errors, transcription errors and mistakes in a document submitted to the EPO to be accepted, the following requirements must be met (G 1/12):
1) The correction must introduce what was intended from the beginning. The possibility of correction cannot be used to allow a person to change his or her mind or refine his or her plans. It is the actual intention of the parties that must be taken into account, not their ostensible intention.
(2) Where the original intent is not immediately apparent, the claimant bears the burden of proof, and that burden must be high.
③The error subject to correction may be a misstatement or an omission.
④A request for correction must be submitted without undue delay.
For further information on the EPC Guidelines, please see the following pages:
https://www.epo.org/en/legal/guidelines-epc