By Christian Heubeck and Wolfgang Weiß, Weickmann & Weickmann.This article first appeared in IAM Yearbook 2021, a supplement to IAM, published by Globe Business Media Group - IP Division.
In the fields of molecular biology, pharmaceutics, diagnostics and medical technology, great efforts have been made in research and development for strategies to fight covid-19. Many of these efforts are focused on new diagnostic methods, medicaments and vaccines, as well as medical devices.
New inventions in these fields may, of course, be protected by patents, thereby providing a monopoly for the originators and prohibiting competitors of entering the market. Under the European patent system, most of these innovations are subject to patent protection. This applies for novel chemical or biological compounds, including polypeptides such as antibodies or viral antigens and nucleic acid molecules, irrespective of whether these agents might be found in nature. In addition, patent protection may be obtained for novel therapeutic applications of known compounds and compositions. According to EPO case law, any novel pharmaceutical use must be made plausible in the patent application (eg, by experimental data). In case such experimental data is unavailable, plausibility might be established by including a detailed study protocol of the claimed medical use.
In the diagnostic field, *in vitro* methods may be patented. For *in vivo* diagnostic methods, a different claim format directed at the diagnostic agent for in an *in vivo* diagnostic method must be selected.
In the field of medical devices, however, there is a protection gap in the European patent system. While a medical device may be patented as such (ie, as a product), a novel use or method involving a known medical device cannot be patented via the European patent system.
During the coronavirus pandemic, serious concerns have arisen among the public regarding whether the existing patent system is adequate to meet the needs of providing protection against covid-19 for the whole population at reasonable costs. There are particular concerns that medicaments or vaccines against the coronavirus might be monopolised by patent owners and limited to those who can afford to pay a very high price.
While these concerns might not be completely unfounded, we believe that they will play no major role, at least in the near future. First, the EU patent system provides the legal option of suing patent owners to grant a compulsory licence under a patent to third parties if they cannot meet an urgent need for supplying the protected products, applications or methods to the public. Second, in recent months, many countries have passed emergency legislation allowing governments direct access to patented technologies in case of public need. Third, in the age of social media, patent owners are very careful to avoid giving the impression that, for financial interests, they will block the public’s access to greatly needed patented technologies. The present European patent system should be well equipped to handle the coronavirus crisis without the need of major revisions.
IP protection of AI-based inventions
As in most industries, AI is set to play an increasing role in the life sciences and pharmaceutical industries. Presently, following transport and telecoms, life sciences and pharmaceutical patent applications represent the third largest group of AI-based patent applications.
In general, the pharmaceutical industry uses AI in research and development for analysing big data. Such AI-derived solutions are becoming increasingly important when it comes to significantly reducing time and costs. AI can process vast amounts of data, detecting and interpreting patterns that were previously impossible to identify or even imagine. There is a wide variety of possible uses. For example, AI may be used for drug-target identification and validation, phenotypic drug discovery, identification of biomarkers, polypharmacology discovery and drug repurposing, among others. Such AI-based approaches are also closely interrelated to another promising and highly relevant field of pharmacologic research – personalised medicines. Do AI-based methods for analysing patents, patent applications, publications and research literature provide for a strong promising application?
Of course, the results of AI-based methods can often, in themselves, be the subject of patent applications, such as newly identified drugs and biomarkers. The basic principles for patenting AI-based methods themselves are still in a discovery phase, but the EPO has provided guidelines for applications directed at AI-based methods in its Guidelines for Examination.
According to the EPO Guidelines for Examination, AI is based on computational models and algorithms for classification, clustering, regression and dimensionality reduction, such as 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’ based on training data (Guidelines G-II, 3.3.1). The general guidance provided for mathematical methods (G-II, 3.3) also applies to AI-based methods.
According to Articles 52(2) and (3) of the European Patent Convention (EPC), no European patent will be granted for mathematical methods, mental acts, computer programs or presentation of information, among other things, as they are not regarded as inventions. However, if a claim is directed either at a method involving the use of technical means (eg, a computer, processor or storage media) or at a device, its subject matter has a technical character as a whole and is therefore not excluded from patentability. Technical character of a feature results either from the physical features of an entity or (for a method) from the use of technical means. General terms such as ‘support vector machine’ or ‘neural network’ may, depending on the context, merely refer to abstract models or algorithms devoid of technical character and do not, on their own, necessarily imply the use of a technical means (G-II, 3.3.1).
Merely specifying the technical nature of the data or parameters of the mathematical method may not be sufficient to define an invention in the sense of Article 52(1) of the EPC, as the resulting method may still fall under the excluded category of methods for performing mental acts as such (Articles 52(2)(c) and (3); see G-II, 3.5.1).
It is legitimate to have a mix of technical and non-technical features appearing in a claim, in which the non-technical features may even form the predominant part of the claimed subject matter (T 154/04; T 1784/06). Novelty and inventive step, however, can only be based on technical features, which must be clearly defined in the claim. Non-technical features, to the extent that they do not interact with the technical subject matter of the claim for solving a technical problem (ie, non-technical features ‘as such’), do not provide a technical contribution to the prior art and are therefore ignored in assessing novelty and inventive step (T 154/04). If non-technical features belong to the general framework in which the invention evolves, they may be used in formulating the relevant technical problem (T 641/00; T 77/14) for assessing inventive step using the problem-solution approach. Therefore, in claims comprising technical and non-technical features it must be evaluated for each feature, in the context of the invention, if it contributes to the technical character of the invention, since this is relevant for assessing inventive step and novelty (G-II, 2).
Once it is established that the claimed subject matter as a whole is not excluded from patentability under Articles 52(2) and (3) of the EPC (ie, it comprises technical features and is therefore an invention in the sense of Article 52(1)), it is examined in respect of the other requirements of patentability, in particular, novelty and inventive step (G‑I, 1). For the assessment of inventive step, all features which contribute to the technical character of the invention must be taken into account (G‑VII, 5.4).
Like mathematical methods, AI-based methods may contribute to the technical character of an invention (ie, contribute to producing a technical effect that serves a technical purpose, by its application to a field of technology and by being adapted to a specific technical implementation (T 2330/13)). When assessing the contribution made by an AI-based method to the technical character of an invention, it must be taken into account whether the method, in the context of the invention, serves a technical purpose (T 1227/05, T 1358/09). According to the guidelines, examples of technical purposes which may be served by an AI-based method in the life sciences field are “providing a genotype estimate based on an analysis of DNA samples, as well as providing a confidence interval for this estimate so as to quantify its reliability” or “providing a medical diagnosis by an automated system processing physiological measurements”.
Generic purposes such as ‘controlling a technical system’ will not be accepted. The mere fact that an AI-based method may serve a technical purpose is not sufficient either. The technical purpose must be specific. The claim is to be functionally limited to the technical purpose, either explicitly or implicitly.
Further, like a mathematical method, an AI-based method may also contribute to the technical character of the invention independently of any technical application when the claim is directed to a specific technical implementation of the AI-based method, and the AI-based method is particularly adapted for that implementation in that its design is motivated by technical considerations of the internal functioning of the computer (T 1358/09). If the AI-based method does not serve a technical purpose and the claimed technical implementation does not go beyond a generic technical implementation, the AI-based method does not contribute to the technical character of the invention. In such a case, it is not sufficient that the AI-based method is algorithmically more efficient than prior art AI-based methods to establish a technical effect (see also G‑II, 3.6). However, if it is established that the AI-based method produces a technical effect when it is applied to a field of technology and/or adapted to a specific technical implementation, the computational efficiency of the steps affecting that established technical effect should be considered when assessing inventive step.
AI-based inventions and methods can be considered almost as a subgroup of computer-implemented inventions. This leads to the following considerations when drafting patent applications:
In this context, it is highly recommended that the description is drawn up with as much detail as possible, since, for example, the definition of the ‘person skilled in the art’ in the fast-developing AI field and, therefore, its consequences for the assessment of inventive step, is currently under discussion and has not yet been finally clarified.
Regarding AI-based property rights, in addition to the patenting requirements of novelty and inventive step, a number of other questions have arisen that have not yet been clarified. For example, with regard to sufficiency of disclosure or enablement it has not been decided whether, and if, how detailed the source code of an AI-based method must be disclosed in the corresponding application to meet the requirements of Article 83 of the EPC (ie, disclose the invention in a manner sufficiently clear and complete for it to be carried out by a person skilled in the art).
AI as inventor
It was recently decided in EPO proceedings that an AI system cannot be an inventor. In January 2020, the EPO published its decisions setting out the reasons for its refusal of the two European patent applications EP 18 275 163 and EP 18 275 174 in which an AI system was designated as the inventor (see www.epo.org/newsevents/news/2020/20200128.html).
In both applications a machine called DABUS, which is described as “a type of connectionist artificial intelligence”, is named as the inventor. The applicant stated that he had acquired the right to the European patent from the inventor by being its successor in title, arguing that as the machine’s owner, he was assigned any IP rights created by this machine.
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 EPO 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. Moreover, the designation of an inventor is mandatory, as it bears a series of legal consequences, notably to ensure that the designated inventor is the legitimate one and that he or she can benefit from rights linked to this status. To exercise these rights, the inventor must have a legal personality that AI systems or machines do not enjoy.
Both decisions are under appeal.
As the above case illustrates, there is still very little case law from the EPO Boards of Appeal regarding the patentability requirements for AI-based inventions. It will probably take several years before an established case law is developed.
In the meantime, however, there are a number of other protection options for AI-based inventions besides patent applications, which applicants can consider when building an IP portfolio. Apart from copyright protection of, for example, source code or database structures written by programmers, trade secrets should also be considered. Directive 2016/943 of the European Parliament harmonised the protection of undisclosed know-how and business information (trade secrets) against their unlawful acquisition, use and disclosure across the European Union.