China and the US lead in patents and publications in the field of generative AI
Over the past 10 years, the number of registered patents in the GenAI field has increased from 733 in 2014 to more than 14,000 in 2023. Since 2017, there has been high dynamics of patent activity – an average annual growth of about 45%. This is due to the fact that in 2017, Google researchers published a breakthrough scientific publication on transformers (this concept made it possible to create advanced large language models, such as ChatGPT).
The number of scientific publications in the GenAI field has grown by two orders of magnitude over the same period, from just 116 to over 34,000. 2023 saw an explosive growth in scientific publications, tripling in just one year.
There was no acceleration in patent growth in 2023. However, it can be expected in 2024 and 2025, as there is usually an 18-month lag between patents and scientific publications.
It is worth noting that GenAI patents represent only a small share of the total number of patents related to artificial intelligence (AI). In 2023, there were 14.1 thousand GenAI patents, compared to almost 230 thousand AI patents overall.
China is the world leader in GenAI patents by a wide margin. From 2014 to 2023, China accounted for 70% of the world's patent share, and this share has increased year after year. Chinese companies and researchers registered 38.2 thousand patents, while American companies registered only 6.3 thousand. Korea (4.2 thousand) and Japan (3.4 thousand) follow, while India (1.3 thousand) and European countries (hundreds of patents for Britain and Germany) lag significantly behind.
The leading companies are China's Tencent Holdings (2,074 patents), Ping An Insurance (1,564), Baidu (1,234), Alibaba Group (571), ByteDance (418), the American IBM (601), Alphabet/Google (443), Microsoft (377), and the Korean Samsung (468).
In terms of scientific publications, China and the United States dominate to a large extent, being roughly equal in terms of publication numbers, with a slight advantage for China.
However, the two countries differ greatly in citations (citations to publications in scientific papers), and citations are a measure of paper quality (albeit an imperfect one). Scientific publications with at least one author from the United States received significantly more citations overall than those with at least one author from China.
The ranking of research organizations by number of scientific publications is dominated by Chinese and American institutions, with several positions in the top 20 (China has eight, the US has six). In addition, there are four universities from the UK in the top 20, and one each from Canada and Japan. The Chinese Academy of Sciences is clearly ahead with more than 1,100 scientific publications. Tsinghua University and Stanford University follow in 2nd and 3rd positions, with more than 600 scientific publications each. Alphabet/Google (fourth place) is the only for-profit company in the top 20 institutions, with 556 scientific publications.
The number of citations of scientific publications shows a higher position for American companies. Alphabet/Google is the leading research organization in this ranking by a large margin, and seven other American commercial companies are present in the top 20. The Chinese Academy of Sciences is in ninth place, behind even a Korean and Canadian university.
The report concludes that we are only seeing the beginning of a wave of GenAI patents. The World Intellectual Property Organization also cites Goldman Sachs’ estimate that more than 300 million jobs worldwide “will be subject to some degree of automation.”
P.S. All rights reserved by WIPO. Copying and distribution of the report is encouraged.
