Valohai blog

Insights from the deep learning industry.

All Posts

5 Interesting Things About AI and Patenting

All over the world, patents are known as the best way to protect inventions. They provide inventors with a period of up to 20 years to use an exclusive, monopoly-like position in the commercial exploitation of their creations. It is the key for getting returns on the investments they made during the research and development of their new technological solutions.

With billions of dollars invested in the AI industry, the need for patent protection is self-evident. However, do the inventions in the field of AI fit well into the legal framework designed in the pre-internet era? Or is there anything in the current patent system that has to be adapted to the new technological reality? We are going to talk about this during our Fireside Chat in a few weeks, but before that, here are five interesting facts about AI and Patenting.

  1. Abstract mathematical methods cannot be patented if claimed as such. AI and Machine Learning are based on computational models and algorithms, meaning that they have an abstract mathematical nature. According to patent law, they cannot be patented if claimed as such, whatever unique and valuable they are. Patenting requires a clear technological character, which we will discuss in more detail in our chat.
  2. Simple automation of functions previously performed by humans is not considered inventive. By way of example, think about self-driving cars. While the AI-based programming part, that enables the car to drive itself without human intervention, can be protected by patents, the whole concept of self-driving car does not get that protection.
  3. Following the path of software patenting. Patenting in the field of AI began long before the recent breakthrough in Deep Neural Networks. No special regulation was needed at that time to facilitate AI patenting. The process is similar to software patenting, although all the issues in case of AI would be magnified.
  4. Computer Vision attracts the most attention from developers. According to the latest State of AI Report, computer vision is the most popular area for patents; and the leading number of applications comes from the field of biometrics.
  5. Can AI be an inventor? AI models can be used as a tool in the invention process. For example, the Oral-b CrossAction toothbrush and the Airbus 3-D printed airplane partitioning wall, to name two, were invented in this way. However, should we allow patenting something that was invented not by humans? This question is capable of breaking the foundations of the global IP legal system and no country in the world has come up with a workable solution yet.
Are you interested to hear more about these and other aspects of patenting AI solutions?

Sign up for our Fireside Chat on August 15th.
Vadym Kublik
Vadym Kublik
Information Technology Law Expert

Related Posts

5 Interesting Things About AI and Patenting

All over the world, patents are known as the best way to protect inventions. They provide inventors with a period of up to 20 years to use an exclusive, monopoly-like position in the commercial exploitation of their creations. It is the key for getting returns on the investments they made during the research and development of their new technological solutions.

Watch the Webinar on Version Control in Machine Learning

Watch a recording of the webinar on version control in machine learning that was held on 22th of November 2018. During the webinar we discussed about the topics below and answered multiple questions addressed by the attendees.