Home / Autre / Tout le monde parle, peu rapportent : En 2023, 14 000 brevets GenAI ont été déposés

Tout le monde parle, peu rapportent : En 2023, 14 000 brevets GenAI ont été déposés

No matter how smart it is, artificial intelligence would not exist without its inventors who diligently work on many discoveries, as evidenced by the large number of GenAI (generative artificial intelligence) patents, which have increased by more than 800 percent over the past decade. According to data from the World Intellectual Property Organization (WIPO), there were only 733 patents in 2014, and in 2023, there are as many as 14 thousand. During the same period, the number of scientific publications on artificial intelligence has increased even more: in 2014, only 116 were published, while in 2023, more than 34 thousand were published, meaning that just last year, more than 25 percent of all GenAI patents were granted and more than 45 percent of all GenAI scientific papers were published. Inventions in artificial intelligence are based on computational models and algorithms such as neural networks, genetic algorithms, machine learning algorithms… These models and algorithms are of a mathematical nature, using large input data and providing various outputs.

The rise of GenAI in recent years has been primarily driven by more powerful computers, the availability of large datasets as data sources, and improved artificial intelligence and machine learning algorithms. For example, the development of transformer architecture in LLM has significantly advanced GenAI. The development of complex applications in many different areas has been enabled, but without top inventors, there would be no sudden increase in patents and progress in GenAI. Their revolutionary inventions have not only advanced artificial intelligence (AI) technology but have also made it an indispensable part of many industries in a very short period.

There is hardly a field without it

There is hardly any activity where artificial intelligence is not used. It is even present in art, where the question is increasingly raised whether certain works were created by a human or AI. Among the most important GenAI models are generative adversarial networks (GAN), variational autoencoders (VAE), and decoder-based large language models (LLM). Key application areas for GenAI patents include software, medicine, document management, business solutions, industry, manufacturing, transportation, security, telecommunications…

GenAI is already present in many products, services, and processes; thus, it becomes a technological driver for content creation and efficiency improvement and will further impact many industries in the future. For example, it can accelerate drug development by designing molecules for new drug formulations and personalized medicine. In document management, it can automate tasks, save time and money, and create customized marketing materials, while in business solutions, chatbots can be used for customer service, retail assistance systems, and employee training. In industry and manufacturing, GenAI enables product design optimization and digital programming, while in transportation, it plays a key role in autonomous driving and public transport optimization.

Top inventors

The most successful AI innovator in the world is engineer Jianzong Wang from the Chinese company Ping An Insurance, who has the highest number of both granted and filed patents. He tops the list with a total of 1077 patents and applications. He currently has 364 granted patents, while 713 applications are still pending. He is followed by American engineer Sarbajit K. Rakshit from IBM, who has 356 granted patents, as well as those awaiting approval. Following Rakshit on the list of the most successful American inventors in the field of artificial intelligence are Kenneth Neumann from KPN Innovations and Austin Walters from Capital One Financial Corporation, who has 303 granted patents and pending applications. They are followed by inventors from companies such as Cisco Systems, Adobe, Stradvision, and Samsung, whose artificial intelligence patents have a significant impact across various industries.

The number of invention applications awaiting approval also reveals the great potential of AI solutions, and their inventors are the driving force behind the evolution of the artificial intelligence sector. The potential of artificial intelligence permeates almost all industries, from mobility, manufacturing, and medical technology, translation services to the future of digital communication, and it is expected to help improve energy efficiency and the security of wireless networks. Although 2023 was particularly important for artificial intelligence due to the large number of filed and granted patents, it should not be forgotten that that year saw the first law on artificial intelligence enacted globally, realized by the European Union. The increase in innovations in artificial intelligence is embedded in global innovation ecosystems, and their evolution is reflected in the latest patenting trends. For example, more than two thousand European patents relate solely to artificial neural networks that process information for speech recognition, image analysis, and autonomous robots in a manner inspired by biological neural systems.

The most inventive companies

Public interest in GenAI was significantly increased by the emergence of OpenAI’s ChatGPT chatbot at the end of 2022. Nvidia’s CEO, American engineer and entrepreneur Jen-Hsun Huang, compared the emergence of ChatGPT to the invention of the iPhone. ChatGPT has made it easier for users to access advanced GenAI programs, especially large language models (LLM). These models have reached new levels of performance, demonstrating potential for various real-world applications, triggering a wave of research and development and significant corporate investments in GenAI. On the list of the ten world companies and organizations with the most patents in GenAI are Tencent, a Chinese multimedia company, the largest in the video game sector and with the highest revenue in the world in its sector; followed by Ping An Insurance Group, Baidu, and the Chinese Academy of Sciences

Only Chinese academics

The only research institution on the list with the most patents is the Chinese Academy of Sciences, which ranks fourth; followed by companies IBM, Alibaba, Samsung Electronics, Alphabet, Bytedance, and Microsoft. Tencent, Ping An Insurance Group, IBM, and Baidu hold the highest number of GenAI patents in the world.

Tencent plans to add GenAI capabilities to its products like WeChat to enhance user experience, Ping An Insurance Group is focused on GenAI models for insurance and risk assessment, and Baidu, as one of the first players in GenAI, recently introduced its latest AI chatbot Ernie 4.0 based on LLM. IBM has developed the GenAI platform Watsonx, which allows companies to implement and customize LLMs focused on data security and compliance. Alphabet’s AI division DeepMind recently announced its latest LLM model Gemini, which is gradually being integrated into Alphabet’s products and services. Microsoft is another key player in GenAI and an investor in OpenAI, while OpenAI itself has only recently filed its first GenAI patents.

Published scientists

However, there are no patents without scientific research on GenAI, which also implies publication in scientific journals. The Chinese Academy of Sciences leads in this regard, having published more than 1100 publications since 2010. In second and third place are Tsinghua University and Stanford University; each has published more than 600 publications. Last year, the AI Index from Stanford University reported that existing AI is already helping to build better AI after a leading American chip designer announced the use of machine learning approaches to improve the design of chips that power AI systems.

The largest chip manufacturers in major innovation hubs are now in a race to produce two-nanometer chips, one of the next-generation technologies that will help AI algorithms compute even faster while consuming much less energy. According to records of published scientific publications in companies, Alphabet’s researchers have published more than their competitors, with 556 publications. Alphabet also leads in terms of citations. In this regard, it has surpassed the Chinese Academy of Sciences, which dominates in the number of published scientific publications but not in the number of scientific citations. Additionally, Alphabet leads in relation to all other companies on the list of the twenty most successful by citations, with seven. However, in this story, OpenAI is a special case. Although this company has published only 48 scientific texts on GenAI, which places it at 325th among published publications, it has been cited as many as 11,816 times.

There is data, there is no data

According to the total number of GenAI innovations in the world, China leads with 38,210; followed by the USA with 6,276, South Korea with 4,155, Japan with 3,409, India with 1,350, the United Kingdom with 714, and Germany with 708 patents. These countries hold as much as 94 percent of all GenAI patents in the world. As for the European Union, in 2023, computer technology was the third largest technological area for patent applications at the European Patent Office (EPO) with steady growth over the past decade. The latest EU plans to deepen and expand the innovation ecosystem include providing access to a network of European supercomputers dedicated to artificial intelligence to many public and private users, including startups, small and medium enterprises for training their models.

Additionally, the EPO has created the Deep Tech Finder, which includes startups from across Europe specializing in core artificial intelligence activities. Although 130 Croatian startups are engaged in the development of artificial intelligence, there is currently no data on whether and how many GenAI patents have been filed in Croatia. According to Martina Silov, president of the Croatian Association for Artificial Intelligence (CroAI), such data does not exist. However, this does not mean that patents will not soon exist, as Croatian scientists and engineers are of high quality and are in line with global trends in AI. The problem is that there is a delay in realization, primarily due to a lack of financial support for research and development.

Étiquetté :