MIC compared and cross-examined 22,976 AI-related patents and found that 66.2% of patents are used for Natural Language Data Processing, followed by E Expert Systems(5.4%), Machine Learning(4.2%), Adaptive Control Systems(3.6%), Neural Network(1.2%), Fuzzy Logic(1.1%), Genetic Models(0.3%), and Chaos Models(0.3%).

"It is noteworthy that the patent counts of Machine Learning remain relatively low despite its growing popularity and is worth the investment in resources and time. Based on our findings, three representative vendors Microsoft, IBM, and Google have all concentrated on Natural Language Data Processing," says David Chen, senior industry analyst with MIC.

“IBM has a lot more patent counts in Expert Systems compared to Microsoft and Google and this is maybe got something to do with its cognitive system Watson. Most of Machine Learning patents however fall into the hands of Microsoft, Google and IBM.”