Today IBM announced that for the 16th year in a row, it received more US patents than any other company, The number for 2008 was also a record, 4186, for any company. Here are a couple of links to the story:
- New York Times: “Patent King I.B.M. Will Give Away More Ideas”
- Reuters: “IBM wins most U.S. patents in 2008”
More than than the number, I want to talk about two other aspects of the announcement mentioned in the (rather long) title of the press release “IBM Shatters U.S. Patent Record; Will Openly Publish Many More Future Inventions; IBM Research to Work on Patent Quality Index.” From the release:
IBM used the occasion to announce plans to help stimulate innovation and economic growth. The company plans to increase by 50% — to more than 3,000 — the number of technical inventions it publishes annually instead of seeking patent protection. This will make these inventions freely available to others.
IBM also will contribute the advanced statistical and analytical capabilities of IBM Research to a collaborative project that is developing an empirical measure of patent quality.
IBM researchers will join a project aimed at developing a Patent Quality Index to address the issue of low-quality patents — those with uncertain scope or dubious claims to technological innovation — whose number has increased substantially in recent years, together with historic backlogs, creating uncertainty around intellectual property rights, and spawning increased speculation and litigation.
IBM researchers, building upon work by Professor Ronald J. Mann, Co-Chair, Charles E. Gerber Transactional Studies Program of Columbia Law School and Professor Toshiya Watanabe, Research Center for Advanced Science and Technology, the University of Tokyo, will pursue improved patent quality by applying their advanced statistical and data analytics expertise to help create a Patent Quality Index that can assist patent applicants, examiners and the public to objectively assess the quality of patent applications and issued patents.
The goal of this effort is to improve the patent system by establishing empirical, objective metrics directly correlated to factors such as clarity of claims and quality of prior art cited during patent examination. This will help inventors file better applications and examiners make better decisions more quickly, so that patents are more likely valid.
In the first case, IBM is saying that it will significantly increase the number of technical inventions it formally publishes into the public domain. These then become unpatentable by anyone but it also means they are free for use by anyone in anything, including open source and standards. So just as IBM has frequently talked about having a balanced and dynamic intellectual property strategy, this will more explicitly push more useful ideas into the open.
This continues what IBM has been doing with pledging patents to open source, pledging patents for healthcare and education standards, pledging patents for interoperability standards, and the Eco-Patent Commons.
The last part of the announcement is around patent quality. Given the huge volume of patent applications, are there ways for examiners to automatically get measurements on the likelihood of an application representing new and significant work? Can these techniques be used by technologists and IP attorneys before they submit an application to know the chances of acceptance? In the future, can repositories of published inventions be processed automatically when looking for prior art?
While IBM has spoken about patent quality before, today’s announcement highlights that scientists at IBM Research will join the academic and legal work already underway. Techniques will include statistical analysis, textual metrics, and modeling of differentiating factors between patents.
These are not simple projects and there is much work to be done. What IBM is trying to do is innovate rapidly and successfully, and bring scientific approaches to measuring quality. At the same time, IBM is leading in significant inventions patented but also committing to make significantly more inventions freely available to support open source and standards.