AI Talk: Patents, IP trends and taxes

January 17th, 2020 / By V. “Juggy” Jagannathan, PhD

This week’s AI Talk…

AI and Patent law

This week in MIT Technology Review, I saw an interesting article with the thesis “Can a machine be a holder of a patent?” One company, appropriately named Imagination Engines did just that. They fed an AI algorithm all kinds of data and managed to get a few novel ideas out of the system. They then applied for a patent for the ideas—one related to using fractals in designing containers and another using fractals in designing warning lights! The U.K. and European Union patent offices rejected the patent, stating the patent holder needed to be human! The company is now arguing that the entity that came up with the solution needs to be the holder of the patent—even though, of course, the company is going to benefit from the patent. They are arguing that without changes to the patent law, any invention through their machines becomes unpatentable! They are now trying to do this in the U.S. Perhaps, their concerns will fall on more receptive ears in the U.S., as here corporations are people, as judged by the U.S. Supreme Court!

IP technology trends

The World Intellectual Property Organization (WIPO) just released a report titled Technology Trends 2019. I didn’t have a clue such an organization existed until I saw this report! Well, they exist, and their report has a foreword from none other than the famous AI entrepreneur from Stanford: Andrew Ng! He does raise the red flag regarding taking care of those impacted by AI automation. As the report is from an IP organization, it has lots of data related to patent filing in the area of AI and machine learning (ML). It turns out ML is the most dominant AI technique in patent disclosures. The report has an obligatory section on AI basics, but the trends section is good. There’s a lot of trend data related to patents: how many, which countries, which companies, which universities, which techniques, which business segment and more! Chinese universities are strongly represented in all patent categories related to AI and ML. The U.S. and China have the most patent filings, with Japan coming in a close third, followed by other major developed and developing nations. The report also has a section related to policy considerations on the impact of AI with respect to employment, privacy, security and ethics. It reiterates several themes we have already examined in this blog series. The report also examines issues related to patent protection, litigation and policies. They conclude with a call to action for creating a sort of AI “commons” that will provide open access to basic science and technology. All in all, it’s a good report to review, particularly if you are interested in the intersection of the fields of AI and IP!

Robot tax and unemployment

The preoccupation with what to do when jobs are displaced by automation made it into a recent issue of the Wall Street Journal (WSJ). In this article, they debate taxing robots. The rationale is, as one replaces people with automation, the tax revenue base shrinks. Why not replace it with a tax on the automation? This is an idea endorsed by Bill Gates. The revenue generated would be used to retrain workers. The WSJ article also explored addressing the robot-driven unemployment scenario with a new idea to “outskill” laid-off workers. What in the world is outskill, you may ask—essentially, it means teaching employees skills that match up with another job. So, say six months before the employees are laid off, they are trained on skills that help him/her get the next job. My guess is such retraining is unlikely to happen voluntarily within companies unless the government steps in and mandates such regulations!

Acknowledgement

The WIPO report was sent by my longtime friend and IIT Madras classmate, Krishnan Sadashiv.

I am always looking for feedback and if you would like me to cover a story, please let me know. “See something, say something”! Leave me a comment below or ask a question on my blogger profile page.

V. “Juggy” Jagannathan, PhD, is Director of Research for 3M M*Modal and is an AI Evangelist with four decades of experience in AI and Computer Science research.