This AI is mine: patent wars in Industry 4.0?

All companies try to take advantage of the data, interconnect humans and machines, machines with each other and be able to offer more and better products and services.

It is already obvious that artificial intelligence is going to play a definite role in our present and future economy, the term industry 4.0 has been coined. All companies try to take advantage of the data, interconnect humans and machines, machines with each other and be able to offer more and better products and services.

But, as the King of Sparta said "my border reaches where the tips of my spears" and intellectual property has been put in place, to establish these borders, possibly after we are faced with an industrial war 4.0.

If we adopt this war language we will see where the different battles can occur.

The first comes from the accumulation of patents and the conflicts they generate and will generate.

A quick look at the European Patent Office (EPO) database shows, as of March 1, 2020 very revealing data.

If we introduce the term "artificial intelligence", 152,292 patents appear between those requested and those granted. And if we look for more specific terms such as “Machine Learning” are 312,265, with “Deep learning” 108,811 and about the latest AI fashion, the GAN (Generative adversarial networks) we find 3,017.

No sector is left out of this game. There are 1,529 patents that use the term "Fintech" and in a related field, although a little broader as is the "Blockchain" there are 22,675. Here we have gone from only a few dozen applications in 2012 to 2,300 in 2017, half of them related to payment protocols.

But who applies for so many patents? If we go to WIPO statistics, we will find American and European, Japanese and Korean companies but the number 1 is ... Huawei, closely followed by ZTE.

While we persisted in the topics (piracy, cheap products, mere manufacturers) the Chinese government designed a plan to turn the country into a technological giant, which leads to being a leader in patents.

Although Panasonic is the world champion with 34,081 patents applied for between 1978 and 2018, Huawei has 33,899 and started in 2000. ZTE has 25,746.

Huawei is the largest applicant of 2018, with 5,405 patents published. ZTE has 2,080 patents published in the same year. 60% of both patent packages refer to digital communications.

Therefore, when problems arise with the Chinese companies from the United States, there are not only national security issues but also AI technological leadership, based largely on patents, as well as business secrets. But, in any case, the conflict between innovative companies is served.

In the world of telecommunications there are essential patents (Standard Essential Patents, SEP), thus declared by ETSI, an international organization that also promotes a system for obtaining licenses under reasonable conditions (FRAND, Fair, reasonable and non-discriminatory). But we do not live in a calmed and peaceful scenario but of increasing belligerence, with sentences from European courts that sometimes do not use the same criteria and with those of the United States, Japan and China also in the game.
The extension of this litigious scenario to the entire AI field is obvious and we are awaiting a harmonized framework, despite the efforts of the EU and WIPO. If we do not find a harmonized, fast and global dispute resolution system, we will be engaged in a trench war.

But there is a second battlefield, as AI is creating and will create distortions to the patent system, in relation to ownership and patentability conditions.

We have already had a debate about the so-called software patents or computer-implemented inventions. After years of frantic production of software patents in the United States, the judgments of the Supreme Court “Bilski v. Kappos ”and especially“ Alice Corp. V. CLS Bank ”supposed an earthquake in its innovative industry, saying that not everything was patentable and that the software could be a simple abstract creation without patentable matter.

Meanwhile, Europe acted with more prudence. The European Patent Office (EPO) developed a rational and prudent policy in the admission of so-called “computer-implemented inventions”, which has evolved in its criteria to adapt to new situations.

AI is a part of computational science and therefore what has been learned from the above is that we must seek a technical effect so that the software-based invention is patentable. But AI develops a "super software" and creates new problems for patents.

The development of inventions in the field of AI involves the use of techniques such as machine learning, Deep learning and GAN, mentioned above. The problem is that the more advanced these techniques are, the more "autonomy" the artificial intelligence entity has, either in the generation of new algorithms from which they are supplied or in the selection and use of data. This creates a sort of “black box” and it is difficult to evaluate the classic criteria for granting patents, sufficiency of description and inventive merit.

This last issue relates to another that has been the subject of many comments, the DABUS case or if an AI entity can be designated as an inventor. In Spain (and in many other countries) only a natural person can be an inventor and the EPO seems to have settled this issue in the same direction. But the debate continues because, if it cannot be an inventor, can rights be granted over the invention to the owner of the machine?

And from all of the above, multiple problems arise in research projects and transactions on intangible assets, in addition to rampant litigation about the validity and infringement of patents, responsibilities and much more.

And there is the issue of trade secrets around AI, its role in this field is relevant and it will be more. But that, as they say in the showbiz, is another story.

Javier Fernandez-Lasquetty

Lawyer, specialist in Intellectual Property and Information Technology. Partner of Elzaburu SLP. Professor and Director of IP programs of the IE Law School. Professor of the WIPO Academy and panelist of its Arbitration and Mediation Center. Member of the board of directors of DENAE and LES España & Portugal. Member of the Academic Council of FIDE, director of the annual Congress on Industrial and Intellectual Property and Co-Director of the International Congress on Artificial Intelligence and Intellectual Property.