Preventive patent enforcement by artificial intelligence
Thomas Heinz Meitinger · Journal of Intellectual Property Law & Practice · 2024
A new world would emerge if modern technologies, especially artificial intelligence (AI), were able to effectively prevent infringements of the law. For example, an AI system in a car could utilize real-time driving data and other information recorded by sensors in the cabin (e.g. heart rate, eye movement, etc.) to detect if a driver is incapacitated, safely parking the car without further driver input and preventing the continuation of the journey. But could AI also prevent the infringement of patents, realizing what could be described as preventive patent enforcement? The prerequisite for preventive patent enforcement is the existence of a claim that can be enforced before an infringement is committed. In the current legal framework, patent law already provides patent holders with a prohibition right even if there is only a risk of patent infringement, ie before an infringement has been committed. A prohibition right due to patent law can therefore be applied preventively.1 In addition, a prohibition right arises due to indirect patent infringement. With the legal institution of indirect patent infringement, the legislator wanted to create a possibility to get to the root of the evil of patent infringement. In particular, the intention was to enable the patent holder to avoid a large number of proceedings against direct patent infringers by means of a single infringement action against an indirect patent infringer.2 The legal institution of indirect patent infringement therefore has a distinctly preventive character. However, indirect patent infringement requires a subjective element, namely intent, which is difficult to prove.3 Before a preventive patent enforcement takes place, the validity of the patent to be enforced should be examined. In addition, the possible patent infringement must be reliably verified. An examination of validity includes, in particular, a novelty test and an assessment of inventive step. An AI will fail at least in the examination of inventive step, as this requires consideration of the ‘inducement’ of the person skilled in the art. A person skilled in the art will not simply combine documents in order to arrive at an invention, but will only perform an overall review of documents if they see a reason to do so. This reason arises from the meaning of the prior art, which is apparent to the person skilled in the art on the basis of their technical knowledge and imagination.4 Currently, the ability of AI to check legal validity is generally denied. However, a change in the interpretation of novelty and in particular inventive step according to patent law could open up the possibility for AI to examine the legal validity of a patent.5 At the very least, it is expected that in the next few years an AI will be available that can evaluate a technical teaching for its novelty and inventive step.6 When assessing patent infringement, the meaning of the claims must also be determined by interpretation. An interpretation is again made from the perspective of a person skilled in the art.7 The patent claims are to be interpreted by the skilled person in such a way that illogical and technically nonsensical embodiments are not taken into account. The person skilled in the art should arrive at technically reasonable embodiments by structuring and taking an overall view of the features.8 An AI cannot currently do this. Preventive patent enforcement is therefore possible in principle due to the preventive nature of the prohibition right in patent law. However, an AI cannot currently simulate a person skilled in the art for the purposes of examining the validity of the patent and its scope of protection, thus making it impossible for AI to determine the existence of a potential patent infringement. Preventive patent enforcement with an AI is therefore not currently possible, although further discussion and research are needed to understand whether preventive AI patent enforcement will become a reality in the near future, and whether this should be seen as a positive development for patent law.