The release of OpenAI’s ChatGPT chatbot in November 2022 has greatly increased public enthusiasm for generative AI (GenAI). It has been described by many, including Nvidia CEO Jen-Hsun Huang, as an “iPhone moment” for GenAI. This is because the OpenAI platform has made it easier for all users to access advanced GenAI programs, particularly large language models (LLMs). These models have reached new levels of performance, demonstrating the potential for various real-world applications, triggering a wave of research and development, and large corporate investments in GenAI.
This WIPO Patent Landscape Report provides observations on patenting activity and scientific publications in the field of GenAI and builds on the 2019 WIPO Technology Trends publication on Artificial Intelligence. It aims to shed light on the current technology development, its changing dynamics and the applications in which GenAI technologies are expected to be used. It also identifies key research countries, companies and organizations.
GenAI patent families and scientific publications have increased significantly since 2017
The rise of GenAI over the past few years has been driven primarily by three factors: more powerful computers, the availability of large datasets as a source of training data, and improved AI/machine learning algorithms. Developments such as the transformer architecture in LLMs have significantly advanced GenAI. This has made it possible to develop complex applications in many different fields.
The technological advances in GenAI are reflected by the sharp increase in patenting activity. Over the past 10 years, the number of patent families in GenAI has grown from just only 733 in 2014 to more than 14,000 in 2023. Since the introduction of the transformer in 2017, the deep neural network architecture behind the Large Language Models that have become synonymous with GenAI, the number of GenAI patents has increased by over 800%. The number of scientific publications has increased even more over the same period, from just 116 in 2014 to more than 34,000 in 2023. Over 25% of all GenAI patents and over 45% of all GenAI scientific papers were published in 2023 alone.
Which are the top organizations with the most patents in GenAI?
1. Tencent
2. Ping An Insurance Group
3. Baidu
4. Chinese Academy of Sciences
5. IBM
Tencent, Ping An Insurance Group and Baidu own the most GenAI patents. Tencent plans to add GenAI capabilities to its products such as WeChat to improve the user experience. Ping An focuses on GenAI models for underwriting and risk assessment. Baidu was one of the early players in GenAI and recently unveiled its latest LLM-based AI chatbot, ERNIE 4.0. The Chinese Academy of Sciences (fourth) is the only research organization in the top 10 ranking. Alibaba (sixth) and Bytedance (ninth) are other Chinese companies in the top 10.
IBM (fifth), Alphabet/Google (eighth) and Microsoft (10th) are the top US companies in terms of GenAI patents. IBM has developed a GenAI platform, watsonx, which enables companies to deploy and customize LLMs with a focus on data security and compliance. Alphabet/Google's AI division DeepMind recently released its latest LLM model, Gemini, which is gradually being integrated into Alphabet/Google's products and services. Microsoft is another key player in GenAI and an investor in OpenAI. OpenAI itself has only recently filed its first GenAI patents. Rounding out the top 10 is electronics conglomerate Samsung Electronics (seventh) from the Republic of Korea.
Which institutions published the most scientific publications on GenAI?
The Chinese Academy of Sciences is clearly in the lead in terms of scientific publications with more than 1,100 publications since 2010. Tsinghua University and Stanford University follow in second and third place with more than 600 publications each. Alphabet/Google (fourth) is the only company in the top 20 (556 scientific publications).
However, when measuring the impact of scientific publications by the number of citations, companies dominate. Alphabet/Google is the leading institution by a wide margin, and seven other companies are present in the top 20. The case of OpenAI is also noteworthy. In our GenAI corpus of scientific publications, the company has published only 48 articles (325th institution in terms of number of publications), but these publications have received a total of 11,816 citations from other scientific publications (13th overall).
Where are the most GenAI technologies invented?
1. China
2. United States
3. Republic of Korea
4. Japan
5. India
6. United Kingdom
7. Germany
Inventors based in China were responsible for more than 38,000 patent families between 2014 and 2023, based on the inventor addresses published on patents. Since 2017, China has published more patents in this field each year than all other countries combined.
With around 6,300 patent families between 2014 and 2023, the US is the second most important research location for GenAI patenting. The Asian countries Republic of Korea, Japan and India are other key research locations for GenAI, all ranking in the top 5 countries worldwide (third, fourth and fifth respectively). The United Kingdom is the leading European location (sixth globally), with 714 patents published in the same period. However, Germany is close behind (708 patent families) and has published more GenAI patents than the UK in recent years. These top inventor locations account for the majority (94%) of global patenting activity related to GenAI.
Which GenAI model has the most patents?
In recent years, a number of GenAI programs, or models, have been developed. Among the most important GenAI models are:
1. generative adversarial networks (GANs)
2. variational autoencoders (VAEs)
3. decoder-based large language models (LLMs)
However, not all GenAI patents can be assigned to these three specific core models based on available information from patent abstracts, claims or titles.
Among these GenAI models, most patents belong to GANs. Between 2014 and 2023, there were 9,700 patent families of this model type, with 2,400 patent families published in 2023 alone. VAEs and LLMs are the second and third largest models in terms of patents, with around 1,800 and 1,300 new patent families respectively between 2014 and 2023.
In terms of patent growth, GAN patents show the strongest increase over the past decade. However, this has slowed down recently. In contrast, diffusion models and LLMs show much higher growth rates over the last three years, with the number of patent families for diffusion models increasing from 18 in 2020 to 441 in 2023 and for LLMs increasing from 53 in 2020 to 881 in 2023. The GenAI boom caused by modern chatbots such as ChatGPT has clearly increased research interest in LLMs.
What are the main types of data used in GenAI patents?
The main GenAI data types include:
Image
Video
Speech
Sound
Music
Among the different GenAI modes, or the type of data input and output, most patents belong to the image/video category. Image/video data is particularly important for GANs. Patents involving the processing of text and speech/sound/music are key data types for LLMs. The remaining modes: 3D image models, chemical molecules/genes/proteins and code/software have far fewer patents so far. As with patents related to GenAI core models, some patents cannot be clearly assigned to a specific data type. In addition, some patents are assigned to more than one mode because certain GenAI models, such as multimodal large language models (MLLMs), overcome the limitation of using only one type of data input or output.
Top application areas of GenAI patents
The key application areas for GenAI patents include:
1. Software
2. Life sciences
3. Document management and publishing
4. Business solutions
5. Industry and manufacturing
6. Transportation
7. Security
8. Telecommunications
GenAI is bound to have a significant impact on many industries as it finds its way into products, services and processes, becoming a technological enabler for content creation and productivity improvement. For example, there are many GenAI patents in life sciences (5,346 patent families between 2014 and 2023) and document management and publishing (4,976). Other notable applications with GenAI patents ranging from around 2,000 to around 5,000 over the same period are business solutions, industry and manufacturing, transportation, security and telecommunications.
In the life sciences sector, GenAI can expedite drug development by screening and designing molecules for new drug formulations and personalized medicine. In document management and publishing, GenAI can automate tasks, save time and money, and create tailored marketing materials. In business solutions, GenAI can be used for customer service chatbots, retail assistance systems, and employee knowledge retrieval. In industry and manufacturing, GenAI enables new features like product design optimization and digital twin programming. In transportation, GenAI plays a crucial role in autonomous driving and public transportation optimization.
However, many patent families (around 29,900 patent families between 2014 and 2023) cannot be assigned to a specific application based on the patent abstract, claims or title. These patents are instead included in the category software/other applications.