This blog is part one of a two-part series on patents and AI technology.
Artificial intelligence (AI) and machine learning systems are powerful driving forces in the modern era—defining new frontiers for complex and advanced software services and products.
With significant innovation in the AI space, companies should consider the importance of filing patent applications to protect their valuable advancements in the AI space.
In this blog, we briefly canvass the statistics on the growing number of filed AI applications and granted patents, and also provide a few examples of recent AI patent applications.
Large Increases In Patent Filings
The number of applications filed and granted in the AI space continues to grow exponentially.
The Canadian Intellectual Property Office issued a report noting that fewer than 2,000 AI-related inventions were patented worldwide in 2002. In 2017, that number increased substantially to slightly less than 20,000.1 Since then, the number of granted AI patents continues to significantly increase.
Another report issued by the United States Patent Office notes that between 2002 and 2018, annual AI patent applications in the U.S. grew from 30,000 in 2002 to more than 60,000 in 2018.2 The report further identifies that the top 30 AI patent owners, as of 2018, as being in the information and communications technology sector—including the likes of IBM, Microsoft and Google.3
Example AI Patents and Applications
Innovation in the AI space spans a large variety of industries and fields. AI has been previously applied to provide enhanced medical systems and diagnosis, improved natural language processing, sophisticated autonomous vehicle control, among other example applications. Below are a few examples of patents and applications based on AI:
- Amazon Technologies Inc. – U.S. Patent No. 10,706,450 (Artificial Intelligence System for Generating Intent-Aware Recommendations). Training and using machine learning models to predict user intent from user search queries, and in turn, generating or filtering customized content suggestions for store items.
- Google's DeepMind Technologies Ltd. – U.S. Application No. 17/266,724 (Machine learning for Determining Protein Structures). Using AI to determine a protein’s three-dimensional shape from its amino acid sequence.
- Apple Inc. – U.S. Patent No. 10,043,516 (Intelligence Automated Assistant). An invention like Apple's Siri that uses machine learning models for speech recognition.
- IBM Corp. – U.S. Patent No. 9,536,194 (Method and System for Exploring the Associations Between Drug Side-Effects and Therapeutic Indications). Machine learning models which predict therapeutic indications and side effects from various drug information source, and a visual analytics system to support the interactive exploration of these associations.
In part two of this blog series, we discuss important challenges and key considerations in patenting AI. If you have any questions about AI patents, please reach out to the Bennett Jones Intellectual Property Law group.
1 Processing Artificial Intelligence: Highlighting the Canadian Patent Landscape, "https://www.ic.gc.ca/eic/site/cipointernet-internetopic.nsf/eng/h_wr04776.html"
2 "Inventing AI: Tracing the diffusion of artificial intelligence with U.S. Patents", October 2020 (United States Patent and Trademark Office), at page 2.
3 Ibid, at page 10.
