What is artificial intelligence and how is it addressed by the USPTO, EPO, and CNIPA?

In general, artificial intelligence (AI) is the automation of something that a human can do, or simulate Intelligent human behavior by machine. In other words, AI does what a human can do but with much more data and processing incoming information. Unfortunately, claiming that AI is in compliance with its standard definition is like a US Section 101 eligibility denial. Europe and China have already modernized their patent examination procedures for artificial intelligence. If the United States continues to conduct the current examination of machine intelligence according to the abstraction doctrine under Alice And mayo The framework created by the Supreme Court, will we leave this industry behind?

Artificial intelligence is an umbrella term that includes Four main categoriesInteractive artificial intelligence, limited memory, artificial intelligence, theory of mind, and self-aware artificial intelligence. Interactive AI includes machines that operate only based on current data entered; Its decisions take into account the current situation only. Interactive AI does not make inferences based on the data entered. Examples of interactive AI include spam filters, Netflix viewing recommendations, and computer chess players.

Limited memory AI is able to make decisions based on data from the recent past and is able to improve decision-making processes over time. This is the category in which the vast majority of R&D and patents take place. Examples of limited memory for AI include self-driving vehicles that are able to interpret data from the environment and make automatic adjustments to behavior when necessary.

Machines become more aware in the next two categories – artificial intelligence that can understand human emotions and make decisions based on that understanding in AI. Self-aware AI is the most futuristic – these machines are able to process other people’s mental states and emotions, as well as own them. Think of robots in Wall-E or, even darker, in I Robot.

When the original patent laws were drafted, lawmakers did not anticipate that one day we might have machines with decision-making capabilities that would mirror those of humans. As a result, the United States Patent and Trademark Office (USPTO), the European Patent Office (EPO), and the China National Intellectual Property Administration (CNIPA) all have substantive eligibility limitations regarding the mental processes and patenting tasks a person can perform. performance, particularly what the human brain can perform, which includes processing information and data and making decisions based on said information and data. The idea is to prevent the patent system from being misused in this way. But in light of the new technologies emerging in the field of artificial intelligence, each of these offices has tried to update their screening procedures to try to capture some of this topic at best.

CNIPA prohibits patenting methods of mental activities. Recently, CNIPA . released Exam Guidelines Draft On examining inventions related to improving AI algorithms (such as deep learning, classification, aggregation, and big data processing). When searching for a “technical solution” that could make machine intelligence patentable, CNIPA suggests looking at improvements to algorithms and big data processing, whether the algorithms have a specific technical relationship with the internal structure of a computer system, and/or improvements to hardware computing efficiency or the impact of implementation. CNIPA considers improvements to data storage volume, data transfer rate, and hardware processing speed as evidence of a technical solution required for a patent.

In March of this year, The European Patent Office 2022 Guidelines for Examination into effect, which expressly states the following:[a] A mathematical method may contribute to the technical character of the invention, i.e. contribute to the production of a technical effect that serves a technical purpose, through its application in the field of technology and/or by adapting it to a specific technical implementation. “The EPO goes so far as to state explicitly that mathematical formulas can be patented if they are used in a specific technical implementation. Specific examples of improvements in technical impact include the efficient use of computer storage or network bandwidth. The EPO has published a series of Examples of mathematical formulas that contribute to the demonstration of technical impact.

The United States Patent and Trademark Office has released its latest version guidance On examination back in 2019. The Guidelines strongly emphasized technical improvements to the device or device operation to overcome an unqualified subject directed to the rejection of an abstraction. Notably, technical improvements in the US essentially exclude end-user benefits, which differ from new CNIPA and EPO practices that allow user benefits as a consideration of technical impact. The United States is also unique with our Supreme Court, which sometimes gets involved in patent matters, particularly with Alice And mayo Resolutions, which supersede any type of USAF office directive. The USPTO Handbook was created within the confines of the Abstract Idea/Law of Nature Framework Alice And mayo, so he wasn’t able to get to what CNIPA and EPO guidelines did in setting mathematical formulas to be patentable when executed by a machine, setting big data processing and improving hardware processing speed to be patentable. So, in terms of machine process and machine intelligence check procedures, we are unfortunately a bit behind.

To learn more, watch the latest IP Practice Vlogs, Available here.

Wen Xie's photo