Generative Artificial Intelligence ( ) Patent Landscape Report

List of Key Chapter Titles

1. Core Concepts of Generative Artificial Intelligence
2. Global Patent and Research Status of Generative Artificial Intelligence
3. Patent Trends in Generative Artificial Intelligence Models
4. Patent Trends in Generative Artificial Intelligence Modes
5. Patent Trends in Generative Artificial Intelligence Applications
6. Concerns Related to the Use of Generative Artificial Intelligence
7. Limitations of Patent Analysis and Future Outlook
8. Main Model Types of Generative Artificial Intelligence
9. Main Data Modes of Generative Artificial Intelligence
10. Key Application Areas of Generative Artificial Intelligence
11. Correlation Between Generative AI Models and Modes
12. Correlation Between Core Generative AI Models and Applications

Document Introduction

Following the release of OpenAI's ChatGPT chatbot in November 2022, Generative Artificial Intelligence (GenAI) garnered widespread global attention, hailed as the "iPhone moment" for the field, sparking a wave of global R&D and large-scale corporate investment. As a significant branch of artificial intelligence, generative AI, leveraging more powerful computing capabilities, massive training datasets, and optimized algorithms, demonstrates immense application potential across multiple domains. The explosive growth in its patent activity also reflects the technology's rapid evolution.

This report, compiled by the World Intellectual Property Organization (WIPO), is based on patent data and scientific literature from 2014-2023. It establishes a multi-dimensional analytical framework to systematically dissect the global patent landscape from three core perspectives: model types of generative AI, data modes (input/output types), and application fields. The report covers core models such as Generative Adversarial Networks (GANs), Large Language Models (LLMs), and Variational Autoencoders (VAEs), as well as major data modes like image/video, text, and speech/music. It comprehensively presents the global distribution of technology R&D, top patent holders, and key innovation regions.

The analysis process employed a mixed-methods research approach, combining traditional patent searches with AI-driven classification tools (such as BERT models). Data was sourced from authoritative platforms like the IFI Claims patent database and The Lens academic literature platform to ensure accuracy and comprehensiveness. Through an in-depth examination of 54,358 patent families, the competitive landscape among major innovating countries such as China, the United States, and South Korea was clarified, and the technological focus areas of top patent-holding institutions like Tencent, Ping An Insurance Group, Baidu, and IBM were identified.

Key findings of the report reveal that following the introduction of the Transformer architecture in 2017, the number of generative AI patents grew by over 800%, with patent families exceeding 14,000 in 2023 alone. China dominates in terms of patent quantity, contributing nearly 70% of global patent families and maintaining an average annual growth rate of 50%. Generative Adversarial Networks (GANs) are the model type with the highest number of patents, while image/video is the most prominent data mode application area. At the application level, software, life sciences, and document management & publishing are the three most concentrated fields for patent filings.

Furthermore, the report discusses potential risks associated with generative AI in areas such as intellectual property, the job market, and data bias, along with the development of relevant global regulatory policies. It also points out the challenges faced by patent analysis methods when dealing with rapidly iterating digital technologies. This report provides defense researchers, policy analysts, technology innovation practitioners, and intellectual property professionals with a panoramic view of generative AI technology R&D and patent strategy, offering authoritative reference for strategic decision-making and technology planning.