We already spoke about Artificial intelligence (AI) and its’s impact on lives and we also apoke about questions arising from intellectual property righs with regards to AI. And even though we still don’t know who owns rights to results of AI, we know who owns patents to AI itself-funny, right?
So, thanks to AI we can see daily progress in some areas, AI-related inventions are flourishing and are actually transcending the border between theory and praxis, so consequentially we can see more and more AI-related patent applications (more than half of the submitted patents after 2013 were in some way connected to AI).
In the paragraphs that follow we will observe AI trends in three categories (AI techniques, AI applications and AI in industrial sector), key players in AI and geographical AI-related trends as described by a publication titled WIPO Technology Trends 2019: Artificial Intelligence World Intellectual Property Organization. In a nutshell:
Machine learning (acquiring knowledge based on experience) is the dominant AI technique disclosed in patents, accounting for 40% off all AI related patents. Between 2013 and 2016 machine learning technique increased by 28%, while »fuzzy logic« and »logic programming increased by 16% and 19% respectively. In terms of machine learning, »deep learning« is the most dominant as it showed a 175% increase in the aforementioned three-year period. »Deep learning« is followed by »Multi-tasking« with a 49% increase, while we can also see growth in the field of neural networks i.e. »latent representation« and »unsupervised learning«
Among AI applications “Computer vision” (i.e. an area of computer science, dealing with computer systems able of interpretation and analysis of visual data) is prevalent. Computer vision is mentioned in 49% of all AI related patents. The percentage of submitted patents including computer vision increased by 24% from 2013 to 2016. Inside computer vision “biometrics” (31% growth) and “scene understanding” (28% growth) have seen the largest increase.
Computer vision is followed by “natural language processing” (14% of all patents) with the emphasis on semantics and sentiment analysis, and “speech processing” (13% of all patents) emphasizing speech2speech, speech recognition and speaker recognition.
Despite the fact that above mentioned areas are growing the fastest when it comes to the number of patent submittals, other areas (such as robotics, methods of control) show incredible growth.
AI in industrial sector
Transport (15% of all AI-related patents), telecommunications (15% of all AI-related patents), and life and health sciences (12%) are noteworthy when it comes to AI in industrial sector. Fastest growing industries are transport, agriculture, computing and government – all of them have shown an increase of at least 30% between 2013 and 2016.
The largest boom of the last ten years happened in transport – more than 8700 patent submittals since 2013 account for a third of all applications.
What are the key players in AI-related market?
Among the top 30 subjects which are applying for patents there are three noticeable groups: businesses, universities, and public research institutions.
Businesses account for 26 out of 30 applicants, wherein most of them are active in the field of consumer electronics, telecommunication and automotive industry.
Largest portfolio of AI-related patents is held by IBM (8290 applications), followed by Microsoft (with 5930 applications). Twelve of the top 20 businesses are based in Japan, three in the USA and two in China.
Above mentioned companies are dealing mostly with computer vision, whereas IBM is focused more on »natural language processing«.
The fastest growing area as established before is machine learning, wherein Chinese Academy has a very impressive deep learning-related portfolio, involving a remarkable 235 patent groups.
Universities and public research institutions also follow similar trends.
What about geography?
Judging from the location of patent application we can determine the geographical trends of AI development. First patents were submitted in Japan in the early 80s. In the years following China and USA have surpassed Japan, since 2014 China is the global leader. The patent application in these three countries (Japan, USA and China) account for 78% of all patent applications.
In the recent years we can observe a growing number of international applications – e.g. PCT system, which enables applicants to file for patent in more than one jurisdiction – which is the cause and consequence of the mentioned trend.
However, overall, we can claim China and USA are global leaders.
So, main players in many areas will also remain main owners of AI…wonder where that will take us.