The question
It is already a fact that AI assistants are widely used for various tasks related to patent drafting, responding to office actions, translations, and even examination by patent offices. It is also a fact that there is a high degree of variability in the success rate of using AI assistants for all the aforementioned tasks.
One could now ask the question: what about the applicant’s strategy with regard to patent protection? Would AI assistants help, or do more harm than good?
The response to the question must start by clarifying briefly what patent strategy is.
General definition of patent strategy
Generally, the term ‘patent strategy’ encompasses decisions that can be grouped under the following two broad categories:
When the applicant has ideas, how can they be filtered and then grouped for the purpose of patent applications? For this discussion, it is presumed that the ideas refer to technical aspects that are inherently patentable.
For each idea, what is the patenting trajectory? Which is the jurisdiction of the first filing, what will be the timing of the filings, and what countries will be selected?
Each of the two categories is discussed in the following sections.
Selection of ideas and grouping them into patent applications
Many inventors have more than one idea.
Thus, the first step of the patent strategy is filtering all the ideas to select the most promising ones in terms of economic potential and realistic chances of being granted patents, and then grouping the filtered ideas into one or more patent applications. This is sometimes called patent mining.
AI assistants cannot be of great help here, except for drafting an executive summary, which should be revised by humans for the following reasons:
Economic potential is a very broad concept with many variables that cannot currently be put into widely accepted metrics because they involve elements of a subjective nature, such as the degree of risk and the degree of knowledge of the relevant economic sector; and
The likelihood that the idea will mature into patents is not fully quantifiable either, especially when a wide array of territories is involved, because the patent laws are different and there is a high degree of subjective appreciation of the patentability criteria.
Decision on the patenting trajectory
Once the first step has been completed, the applicant must make decisions on the patenting trajectory.
In particular, this means:
Selecting the jurisdiction of the first filing (e.g., a WIPO application or a national or regional application, such as an EPO application); and
Selecting the next steps depending on the first selection, such as a WIPO application and national phases. For a WIPO application, the decision may involve optional procedures such as an international preliminary examination report.
AI assistants cannot be of great help here either, except for making summary reports with the costs of different variations, which should be revised by humans for the following reasons:
The costs of the variations are based on many variables that cannot be controlled. For example, if the international search report contains many objections, the applicant can:
Choose to respond, in which case there are costs for responding; or
Choose not to respond, in which case the applicant takes the risk of receiving office actions in the national phases based on the international search report.
The selection of the jurisdictions depends on the economic interests of the applicant, which, in turn, are based on predictions; for example, in which countries the product is expected to be sold and/or manufactured. The predictions may or may not materialise.
Conclusion on the use of AI for patent strategy
The use of AI assistants for building and enforcing patent strategies is not of great help and, thus, not recommendable.
