A news story seems to break almost every week about the application of AI (and more specifically, machine learning) to addressing scientific, engineering and medical problems. As machine learning algorithms become more user-friendly and readily available (Google's TensorFlow (TM), Microsoft's Cognitive Toolkit and Amazon's Machine Learning on AWS, for example), their usage is proliferating. R&D in almost all industries is looking to AI to provide performance and efficiency benefits.

The excitement over AI has led to large numbers of patent applications in this area. As recently reported by Sarah Le Mesurier, the European Patent Office has issued guidelines, to help address the question as to when a patent application in this field can be considered more than just a mathematical method, which would not be patentable on its own.

These guidelines do not address some other points that are relevant to AI-based inventions. A patent for an AI-based technology must still be new and not obvious, just like any other patent. As the implementation of AI gets easier, a presumption that its use is obvious will grow. Even so, there will still be many non-obvious (and therefore potentially patentable) applications of AI across a wide range of fields, from healthcare to manufacturing. Expect to see many more AI patents in the years to come.