AI has become a ubiquitous term today, both inside and outside the tech landscape. It is not only a functional term that points to the power of technology but is also emotive. The ethos behind AI can be summed up in one word: potential. AI points to the future of what is possible.

Since its conception in 1956, innovators and researchers have published over 1.6 million AI-related scientific publications and filed nearly 340,000 AI-related patent applications. Machine Learning (ML) technology is included in 40% of these patent application claims. The functional applications underlying the majority of these patent applications are natural language processing, speech processing, and image recognition.

What is a patent?

A patent is a federal grant of exclusive right given to an inventor who receives an issued patent claim. It can last up to twenty years.

With a patent grant, United States Patent and Trademark Office (USPTO) confers patent owners with the “right to exclude others from making, using, offering for sale, or selling” the invention in the United States or importing the invention into the US.”

A valid patent bars protection for any subsequent independent invention by another inventor, deeming the later inventor the infringer. It is a powerful form of Intellectual Property (IP) protection that protects either the functionality or the aesthetic properties of an invention.

Is AI patent-eligible?

Any new and useful process, machine, article of manufacture or composition of matter may fall under the scope of subject-matter that qualifies for a utility patent. Since AI software is inherently a process, it can be patent eligible, as long as it meets the USPTO’s five rules for patent eligibility.

The rules are as follows: (1) The invention must be a process, machine, or object; (2) the invention must have utility; (3) the invention must be novel or new; (4) the invention must be non-obvious; and (5) the invention must not have been disclosed to the public before the patent application.

To qualify under ‘process, machine, or object’, says JD Houvener, USPTO-licensed attorney with Bold Patents, the subject-matter of the proposed AI patent application claim must cover a unique software invention that is either tied to a machine or that offers an identifiable improvement to society that humans can’t do alone. Where AI requires a computer, processor, or software to analyze or receive data, it is inherently a method with a practical application.

AI represents the ability of a computer to conduct, improve, or manage activities that humans can’t do alone. The sheer weight of that realization is incredible and points back to the concept of the vast potential of this emerging technology.

Most AI technology that is connected to a computer will be patent-eligible. If an invention qualifies, patent protection is an absolute must for inventors because it is such an evolving technological sector and the pace of change is fast. As such, the last thing an inventor wants is to miss out on ownership or exclusivity over what he or she developed.

AI Patent Landscape

International Data Corporation forecasted in 2017 that the AI industry will generate more than $57 billion in 2021. Efforts to patent AI inventions have exponentially increased, seen in how over 9,000 USPTO patent applications in AI-related areas were published in the last four years. There have been over 154,000 global AI patent applications filed since 2010, the majority of which are in the fields of health and digital security.

Microsoft and IBM are at the top of the pack in the AI patent landscape. Tech giant IBM leads with nearly 9,000 AI-related patent applications, followed by Microsoft with close to 6,000.

Effects of AI Patents

If the USPTO system does its job correctly, AI-related issued and pending patents will have a positive impact on the market. As new discoveries are made and new patents are issued, there will be more momentum for innovation which will better society in the long term.

In the short term, however, the USPTO system may reward the larger companies such as IBM and Microsoft rather than scrappy startups. This is because patents are expensive, time-consuming, and require resources that most new companies often don’t have the ability to use on IP protection.

However, more innovation by larger companies may reward the smaller and more agile companies in the long term. This would depend on whether smaller companies are reading and learning from issued patents, so the responsibility is upon them to watch for trends and use the dynamic force of AI in their research.

This article is a contribution of Carly Klein

About the author

Carly Klein is a first-year student at Loyola Law School. A Los Angeles native and a graduate from Boston University with a B.A. in Political Science & Philosophy, she seeks to pursue a career in civil litigation.