AI Patents: What Patent Mapping Can Tell Us About Future AI Technologies

Artificial intelligence (AI) is evolving all the time thanks to the growing quantities of learning data available and advances in affordable access to high computing power but is this technology being adopted equally in all countries? Business Intelligence Consultant Amira Tiga examines AI patent data in Europe, with a focus on France, to uncover what patent mapping can tell us about the state of AI-based innovation.  

Artificial intelligence (AI) is driving major advances in technology and business across a wide range of sectors. With an impact on virtually every aspect of creation, AI-based technology is evolving all the time thanks to the growing quantities of learning data available and advances in affordable access to high computing power. Through patent mapping, we can uncover the true picture of innovation in this fast-moving field, from the main players to the most competitive technology fields.  

AI Patents: The Basics 

What Is Artificial Intelligence? 

In general terms, artificial intelligence (AI) is a branch of computer science that aims to develop machines and systems capable of performing tasks that require human intelligence. Machine learning and deep learning are both applications of AI. With the emergence of new neural network techniques and software in recent years, AI is now widely perceived to be synonymous with "supervised deep machine learning".  

Is AI Protected By IP? 

AI software can be protected by a patent if it complies with the same constraints as conventional software, i.e., if it has a technical character. Technical character is present when the software produces an additional technical effect, i.e., a technical effect that goes beyond the "normal" physical interaction between a computer program (software) and the computer (hardware) on which it runs.  

What is patentable is generally the AI used in a particular technical context to solve a technical problem. What enables the technical effect to be obtained may be linked to the training, the input or output data, etc. 

AI Patent Mapping in Europe from 2010

Patent mapping analysis is conducted using Orbit Intelligence, a leading intellectual property intelligence software dedicated to patent search and analysis. The platform provides access to the world's largest database of patents and scientific literature. For the basis of this research, we used several equations and classes to construct the database, searching as far back as 2010. The study yielded the following results: 

Legal Status of AI Patent Families in Europe  

Numerous patent families were retrieved disclosing artificial intelligence in Europe: 

• 54% of the inventions selected are granted titles,  
• 6% of patent families have lapsed, 
• 39% of the patent families are still alive. 

Patent families are a group of national (member) applications and patents for the same invention. 

Patent Filing Dynamics 

Examination of filing dynamics reveals a spike in applications between 2018 and 2021, showing a strong growth trend during the last five years. This growth is correlated with the interest in AI thanks to major technological advances, particularly in the field of machine learning, which are extending the capabilities of computers and boosting their performance in a wide range of areas (language processing, speech understanding, image recognition, robotics, etc.). These advances open vast prospects for technological innovations. 

Patent Families By Assignee 

Among the top 20 players can be found Robert Bosch, which has a substantial number of AI-related patent families, indicating a strong focus on innovation and IP in this field. Siemens also demonstrates a significant commitment to AI-related innovations with many patent families. This suggests an active engagement in research and development (R&D) in the AI domain, covering a wide spectrum of applications and technologies. Samsung Electronics, a major player in the electronics industry, also has a considerable number of AI-related patent families. This reflects its involvement in AI technologies across various products and services. Likewise, Alphabet, the parent company of Google, has a substantial number of AI-related patent families. This aligns with Google's focus on AI-driven technologies, including machine learning, natural language processing, and other AI applications. Microsoft's significant number of AI-related patent families indicates a strong focus on AI R&D.  

In general, higher numbers of AI-related patent families suggest more extensive involvement and a broader scope of AI-related technologies for each assignee. These numbers underscore the importance of AI innovation across various industries, from electronics and technology to healthcare and automotive. 

Filing Dynamics By Assignee 

Each company's patent filing trends indicate varying strategies and levels of activity in the AI domain. Peaks, troughs, and fluctuations in filing numbers may reflect shifts in research focus, technological advancements, or changes in business strategies. The filings of Robert Bosch were relatively low in the early years, for example, with a notable increase starting around 2015. The peak is in 2020 (557 filings) suggesting a significant focus on AI innovations during that period. Siemens has demonstrated a steady increase in AI patent filings, with a notable peak in 2020 (372 filings). The trend indicates a consistent commitment to AI-related R&D, with a slight decline in the last couple of years.  

Samsung Electronics shows a fluctuating trend in AI patent filings. There was a noticeable increase in filings from 2012 to 2015, followed by a decline in 2016. The filings increased again in 2017 and continued to fluctuate with a peak in 2020 (356 filings). Alphabet's AI patent filings have been steadily increasing over the years, with a significant jump in 2015. The peak is in 2019 (378 filings), indicating a strong emphasis on AI R&D during that period. Microsoft's AI patent filings also exhibit a generally increasing trend, with a peak in 2019 (299 filings). There is a decrease in filings in 2020 and 2021, followed by a slight increase in 2022 and a sharp decline in 2023. Huawei's AI patent filings show a fluctuating pattern, with a notable increase in 2016 and a peak in 2019 (487 filings). There is a decrease in filings in 2020, followed by a slight increase in 2021 and a significant drop in 2022 and 2023. 

Technology Domain By Assignee 

The distribution of patent families across different technology domains for each assignee provides insights into their specific areas of focus and expertise within the broader field of artificial intelligence. Robert Bosch's strong presence in computer technology, control, transport, and measurement domains reflects a diversified commitment to AI applications. The emphasis on transport suggests a focus on AI innovations in the automotive sector, aligning with Bosch's expertise in mobility solutions. The distribution across various domains indicates a comprehensive approach to integrating AI across different technological aspects, from precision measurement to computer-driven control systems.  

Siemens demonstrates a comprehensive engagement in AI across multiple domains, with notable strengths in medical technology, computer technology, and IT methods for management. The focus on medical technology suggests Siemens' commitment to AI-driven innovations in healthcare and diagnostics. Samsung Electronics exhibits a broad spectrum of AI applications, particularly in computer technology, audio-visual technology, and digital communication. The emphasis on audio-visual technology aligns with Samsung's reputation in consumer electronics, indicating AI-driven innovations in image and multimedia processing. The significant presence in computer technology suggests a commitment to AI integration across various products, reflecting Samsung's role as a major player in the electronics industry.  

Alphabet showcases a robust presence in computer technology, search, and IT methods for management. The emphasis on computer technology likely includes AI applications in Google's search algorithms and other software innovations. The distribution across different domains underscores Alphabet's diverse interests in AI, from information management to advancements in digital communication, aligning with the company's overarching role in technology and innovation.  

Finally, Microsoft's strength lies in AI applications across computer technology, electrical machinery, digital communication, and IT methods for management. The notable emphasis on computer technology indicates a broad spectrum of AI-driven innovations in both software and hardware. The focus on electrical machinery suggests Microsoft's interest in AI applications in energy systems. 

AI Patent Mapping in Europe – Key Takeaways 

Our AI patent mapping analysis in Europe and France illustrates a thriving landscape of innovation and competition. Major global players, as well as prominent French institutions and companies, actively contribute to the development of AI technologies. The trends in AI patent filings and the distribution across technology domains highlight the multifaceted and dynamic nature of AI research, with applications spanning from telecommunications to automotive and aerospace industries.  

This mapping provides valuable insights for policymakers, researchers, and industry leaders, emphasizing the importance of ongoing support and collaboration in shaping the future of AI in Europe and France. As AI continues to evolve, these patent trends serve as a roadmap for understanding the technological landscape and identifying opportunities for further research and development. 

For more results from our AI patent mapping, including analysis in additional countries in Europe and further afield, contact Questel’s specialist IP Consulting team.   

About the author    
  
Amira Tiga is a Business Intelligence Consultant at Questel, specializing in IP topics.