Artificial Intelligence Related Patents at the European Patent Office

The growing interest in disruptive technologies such as Artificial Intelligence (AI) is evident. As we have been analyzing, this topic has been increasingly important as inventions achieve new and inspiring results. This article aims to analyze whether there has been a growing trend in AI-related patent applications at the European Patent Office or whether, despite all the euphoria, we are still far from a particularly inventive AI technology boom. Throughout the article, we will analyze some graphs and tables, taking into account so that we can draw some considerations about the technological advance involving AI.

Research Methodology

Before we proceed, we must pay attention to our research methodology, which was based on the following topics:

  1. Search in the Espacenet database of European patent applications containing at least one subgroup of the Cooperative Patent Classification (CPC) mentioned in Table 1 below and filed from 2010 to 2020;
  2. Export of the results containing the publication numbers of European patent applications to EP Bulletin Search and EP Full-Text Search databases;
  3. Using the statistical tools EP Bulletin Search and EP Full-Text Search to obtain the results shown in Tables 1 and 2, and Figures 1 to 3.


Number of Applications vs. Applicants

The data collection shows that there has been and is an exponential increase in the number of European patent applications in AI-related CPC subgroups since 2010, with a current maximum of 975 applications in 2020, as shown in Figure 1 below.


However, the number (975) for 2020 is expected to become even higher, given the fact that there are still secret applications that have not yet been made public through their publication.

Between 2010 and 2020, 1571 AI-related patent applications were filed with the European Patent Office.

Of this total, 9 were refused, 196 were granted, and 1338 are still pending a decision (with a first examination report issued on 59 of these) (see Figure 2).


We can see that there is a high number of pending applications, which is due to the jump growth of applications occurring in the recent years 2019 / 2020.  We can also see that the average (arithmetic mean) time for the Applicant to receive a first examination report is 1035 days, approximately 3 years, and for the Applicant to be notified of the intention to grant a European patent on its application is 1247 days, approximately 3,5 years (see Table 2).

 

Moreover, it is not surprising that large multinational companies dominate the number of AI-related applications.

As Figure 3 shows, Microsoft and Samsung are leading the list of the Top 30 most important applicants in the AI-related CPC subgroups. According to Table 1, Microsoft ranks 1st with 57 applications, followed by Samsung in 2nd place with 50 applications, and Siemens in 3rd place with 44 applications.

Interestingly, Apple is not in the Top 30, while Google is 5th with 36 applications, and Google-owned AI research lab DeepMind is 8th with 16 applications, Huawei is 14th with 9 applications).

Following Siemens in 3rd place is another German company Bosch, ranking in 4th place with 39 applications.

Moreover, German automakers BMW (7 applications) and Volkswagen (6 applications) are also in the Top 30, indicating the impact AI could have on the automotive sector in the near future.



Figure 3: Top 30 Applicants at the European Patent Office for selected (AI-related) CPC subgroups


From the sample collected, it is clear that there has been recently a very significant increase in AI-related patent applications.  Although the numbers are not astronomically surprising, it is possible to see that there is a tendency for them to further multiply.  The major technology players continue to bet on this inventive area, so it is likely that in the medium to long term there will be fierce disputes involving IP assets related to Artificial Intelligence.