Author: Lang Yanyan Time: 03.06.2020
The COVID-19 sweeping the world in 2020 has brought great impact on people’s work and life, and has also pressed the “fast forward” button for global digital change. Before this, most ordinary people may not have a clear concept of “big data” and “digitization”, but this epidemic has made the whole nation feel these “high-tech” firsthand. For example, by using an app for online health check ins, companies and governments can utilize big data for epidemic tracking, thereby better implementing epidemic prevention and control measures. From resuming work and production to cloud office, cloud teaching, and cloud consultation, it can be said that “digital new infrastructure” is the backbone of China’s epidemic prevention war, playing an indelible role in the victory of the war against the epidemic.
With the opening of the two sessions, “new infrastructure” has also been included in the government work report, and various industries have begun to vigorously promote digital transformation, joining the vigorous wave of “new infrastructure”. What is’ new infrastructure ‘? Previously, we understood that infrastructure construction is the construction of railway, highway and other infrastructure, while “new infrastructure” is the construction of new infrastructure in seven fields, including 5G, ultra-high voltage, intercity high-speed railway and urban rail transit, new energy vehicle charging pile, big data center, artificial intelligence, and industrial Internet. Its essence is the infrastructure construction of information digitization.
Figure 1 “New Infrastructure” Diagram
In the construction of information digitization, artificial intelligence and big data, as the underlying core technologies, have also played an important supporting role in several other fields of new infrastructure. And ‘new infrastructure’ highlights the word ‘new’, which naturally leads to the emergence of new technologies. In this situation, the protection of intellectual property rights also faces new development opportunities and challenges. Based on this, this article provides a brief analysis of the intellectual property protection of artificial intelligence and big data in the context of “new infrastructure”.
Firstly, let’s take a look at the trends in patent applications for artificial intelligence and big data. Referring to Figure 2, the overall number of patent applications in the fields of artificial intelligence and big data has been increasing year by year globally. Since 2010, the growth rate of patent applications related to artificial intelligence and big data has significantly accelerated, and the growth rate in recent years has been remarkable (note: some patent data for 2019 and 2020 have not yet been made public). As of 2020, the cumulative number of applications worldwide has exceeded 200000. It can be seen that although the concept of “new infrastructure” is new, Rome was not built in a day. The global patent reserves and layout in the fields of artificial intelligence and big data have already been formed.
Data sourced from Patentcloud
Figure 2: Trends in Patent Applications for Artificial Intelligence and Big Data
Let’s take a look at the intellectual property protection approaches for artificial intelligence and big data separately.
Let’s take a look at the intellectual property protection approaches for artificial intelligence and big data separately.
Figure 3 Schematic diagram of artificial intelligence intellectual property protection path
At present, the hot technologies of artificial intelligence mainly include machine learning, speech recognition, natural language processing, machine vision, biometric recognition, knowledge graph, etc. The most commonly used means of protecting intellectual property rights of artificial intelligence technology are patents and trade secrets. However, given the technological characteristics of the field of artificial intelligence, which involve significant improvements in mathematical models and algorithms, as well as certain peculiarities in intellectual property protection.
On the one hand, algorithms and machine learning models in the field of artificial intelligence belong to the category of rules and methods for intellectual activities, while patent law has clear provisions that rules and methods for intellectual activities cannot be granted patent rights. Therefore, when applying for patents involving algorithms and models, special attention should be paid to the issue of patent protection objects. In view of the development trend of patents in the field of artificial intelligence in recent years, the China National Intellectual Property Administration has also made special amendments to the Guidelines for Patent Examination, which will come into force on February 1, 2020. The revised “Guidelines for Patent Examination” provide clearer examination standards for patents in the field of artificial intelligence: “The examination should be conducted on the solutions that are required to be protected, that is, the solutions defined by the claims. In the examination, technical features and algorithm features or business rules and method features should not be simply separated, but all the contents recorded in the claims should be analyzed as a whole, including the technical means involved, the technical problems solved, and the technical effects obtained.
In practical operation, applicants can apply algorithms or learning models to specific technical fields for patent applications based on industrial mechanisms in specific application areas. By using technical means to solve technical problems and achieve corresponding technical effects, they can avoid the problem of patent protection objects. However, a potential issue that may arise when combined with specific application scenarios is the limitation of the scope of protection. Therefore, it is recommended that applicants try to expand the new application scenarios of algorithms and models as much as possible in order to obtain more comprehensive protection.
On the other hand, for algorithms and learning models, it is difficult to provide evidence of infringement, and it is not easy to obtain it through reverse engineering. If protected by patents, making specific algorithms and models public may not be the best choice for applicants. Protecting the training methods and algorithm models of artificial intelligence systems through trade secrets may be an ideal choice. But in terms of protecting trade secrets, enterprises need to establish strong protection mechanisms to prevent the leakage of trade secrets. For example, management can be strengthened through signing confidentiality agreements, authorizing the circulation of software and algorithm related information, and other means.
On the other hand, artificial intelligence technology involves a lot of software development, and applicants can also protect computer program code by registering copyright to prevent software code from being stolen by others.
Next, let’s take a look at the intellectual property protection methods for big data.
Figure 4 Schematic diagram of intellectual property protection path for big data
At present, the hot technologies related to big data mainly include big data applications, big data platforms, and data sources. Similarly, in terms of protecting forms, it mainly includes several forms such as patents, trade secrets, and copyrights. Like artificial intelligence technology, the protection of big data also has certain particularities.
On the one hand, big data generally involves business methods, specific algorithms, databases, etc., and is also easily attributed to the intellectual activity rules and methods mentioned above, so it does not have patentability. Therefore, enterprises often cannot protect their data arrangement, selection, and calculation with patents. To apply for a patent, one needs to possess technical attributes, as well as novelty, creativity, and practicality. For big data related systems, architectures, and applications that have technical attributes and solve technical problems, patent applications can be considered. On the other hand, original big data sets and program codes selected and arranged from raw data sets can be protected by copyright. On the other hand, data information and processing algorithms related to trade secrets within a company can be protected as trade secrets.
At the same time, enterprises should also pay attention to risk avoidance when it comes to innovation involving big data. Firstly, data collection web crawlers should pay attention to avoiding the risk of copyright infringement; Secondly, in terms of data storage and transmission, attention should be paid to avoiding risks related to technical standards and copyright infringement; Once again, unauthorized access to users’ private data poses a risk of unfair competition.
Finally, whether it is artificial intelligence, big data, or other fields involved in “new infrastructure”, technology innovation enterprises need to conduct comprehensive and professional patent mining, layout, and application through professional intellectual property agents/institutions when protecting intellectual property rights, in order to effectively protect core technologies and demonstrate their capabilities in the wave of “new infrastructure”.
