Remarks by Director Iancu at the Prior Art Archive Launch Event

October 3, 2018

Remarks delivered at the Prior Art Archive Launch Event

Director of the U.S. Patent and Trademark Office Andrei Iancu

October 3, 2018

Cambridge, Massachusetts

As delivered

Good morning everyone. And thank you, Kate (Darling), for that nice introduction. Thank you also to the Massachusetts Institute for Technology and Cisco for co-hosting today’s event and for inviting me to speak.

Here’s a riddle: What do Kuwait, Donald Duck, and our event today in Boston have in common? The answer is, prior art!

On September 14, 1964, the freighter “Al Kuwait” capsized in the fresh-water harbor of Kuwait with a cargo of approximately 5,000 sheep on board. This was a very big deal to the residents of Kuwait because it happened right in the middle of their main source of water, and the idea of having thousands of decomposing carcasses in the water supply would make anyone concerned!

Obviously, the ship had to be raised in order to save the water supply, but nobody was sure how to do it quickly and without causing further damage. Bringing in cranes would have taken too long, and with such methods there is a significant risk that the ship would break. On top of it, a Danish company insured the ship, and they were on the hook for significant loss.

So they asked an inventor named Karl Kroyer for assistance. Kroyer and his team came up with a method of raising the ship by filling it with buoyant balls fed into the ship through a tube. Did this crazy idea work? You bet it did!

After being filled with 27 million polystyrene foam balls—all airlifted from Berlin to Kuwait—the freighter was successfully brought back to the surface. The Kuwaiti water supply was saved.

For his part, Kroyer went on to apply for patents on his method of raising sunken ships, which— on the face of it— certainly seemed commercially viable. Indeed, for this method, he received patents in the U.K. and Germany.

All was going well, until he applied for a Dutch patent. The story is told that Kroyer’s Dutch application was rejected because a piece of previously undiscovered prior art was found that destroyed the novelty of his invention.

So what was this prior art you may ask?

Was it a prior patent? No, this kind of prior art would likely have been discovered in connection with his earlier applications.

Was it a scientific study on buoyancy and water displacement? No. Again, this sort of prior art would likely have been found by prior patent examiners familiar with the subject matter of the claims.

Was it a historical account of prior attempts to raise sunken ships? No. A well-known historical event of such magnitude also may well have been discovered during the earlier examinations of this subject matter.

Instead, it was in fact Disney’s Donald Duck that sunk Kroyer’s patent on how to raise sunken ships.

More specifically, it was a 1949 cartoon called “The Sunken Yacht,” which shows Donald Duck and his nephews—Huey, Dewey, and Louie—raising a sunken ship by filling it with ping-pong balls shoved into the ship with the help of a tube. This Donald Duck cartoon disclosed the very same technique that Kroyer claimed in his patent application. Indeed, the cartoon story mapped out all of the essential elements, and the functions that were carried out matched the functions in the patent application. Consequently, the Donald Duck cartoon was deemed novelty-destroying prior art.

As we know, it doesn’t matter if it comes from an unexpected or hard to discover source; if prior art makes clear that an invention is not new, no patent should issue.

So, in addition to teaching us all how to raise a sunken yacht, it seems Donald Duck has imparted yet another valuable lesson—not just to his waterfowl nephews, but to the patent community at large. And that lesson is that prior art can come from anywhere. It can come from the most unlikely of places. (Who knew Donald Duck could supply prior art?)

And even if the prior art is not surfaced during examination at the Patent Office, it is entirely possible that it will be discovered years later, after issuance, by an opponent during disputes such as litigation or licensing negotiations.

Later-discovered art of this nature tends to introduce a level of uncertainty in the patent system. Our goal, of course, is to reduce uncertainty, and increase predictability and reliability.

Patents are critically important to innovation. They are a driving force of economic development and job creation. But to ensure the patent system operates as intended, it must be reliable, predictable and of high quality.

We all want high quality patents, and I think one of the most important touchpoints of quality is whether an issued patent withstands a fair challenge down the road. Surfacing the most relevant prior art during examination—and examining in light of that prior art—is critical to ensuring this level of quality for the patents that survive the initial examination.

With the most relevant art in hand, one of two things will happen: either we will be able to help the applicant obtain appropriately-scoped claims during the examination, or we will not issue the patent at all. But if we do issue the patent in light of the most relevant art, that patent will have significantly higher odds of success should it be challenged later.

However, uncovering the most relevant prior art during examination has become an increasingly more difficult task; especially with non-patent literature. This is because over the past couple of decades we have been experiencing both a publication explosion and an accessibility explosion. The result is that the amount of published literature has been increasing exponentially. Yet, for any one patent application, there is generally still one examiner with a certain— necessarily limited— number of hours available for examining that application.

On the other hand, parties in litigation or other disputes can more easily scale and dedicate the increased time and resources necessary to find the most relevant art. As a result, a gap can develop between the art identified during examination, and the art an opponent can find during a dispute years later. And the more people publish, the more accentuated the gap could become—unless we do something to change the paradigm.

So in order to increase the reliability of the patent grant, we must work to close this gap. The short of it is that we must work to change the paradigm for finding and surfacing prior art up front, during examination.

That’s why efforts like the Prior Art Archive you are launching today are so incredibly valuable.  When private parties help to collect and surface prior art up front, the resulting examination will be improved, the information gap reduced, and the reliability of the patent grant increased.

Among other things, we need tools to collect the relevant art in concentrated locations in order to reduce the time and resources needed to hunt it down, and tools to search that art efficiently and effectively. This new database is a step in this direction.

Private parties, especially those with long histories of innovation, are often sitting on wellsprings of information that may be highly relevant prior art, but that is also difficult to locate during examination. By collecting this information and making it free, searchable, and readily available to the entire industry, the Prior Art Archive will help our examiners, and the public, to identify the relevant prior art—upfront. Databases such as this one, in other words, can help reduce the prior art gap.  

As Director of the USPTO, one of my highest priorities is to provide our examiners with the tools they need to ensure that our agency issues predictable, reliable and high-quality patent rights. Your new database helps us do that.

At the USPTO, we are also working with a host of other tools and programs to help us in this regard. For example, our Expanded Collaborative Search Pilot is designed to leverage search expertise of examiners from different offices to better identify more comprehensive prior art, particularly foreign prior art, by combining the search results of examiners at the USPTO with the Japan Patent Office and/or the Korean Intellectual Property Office before issuing an office action in the patent application.

Additionally, the USPTO’s Global Dossier provides free access to patent application information from all participating IP offices from a single website, with free English translations. This tool enables our examiners to have streamlined access to work done in foreign offices so they can more easily find relevant references and prior art. We have also established at the USPTO a task force for identifying or creating artificial intelligence (AI) tools to help search the forever increasing amount of information.

And we are in the process of requesting industry input on AI technologies and approaches that can be applied to our patent search and examination activities. In particular, our recent Request for Information, named USPTO’s Challenge to Improve Patent Search with Artificial Intelligence, seeks technology that would, among other things: expand valuable sources of search information, apply emerging technologies to classification of documents, increase the validity and thoroughness of search results, and summarize techniques to determine relevancy. If you have technology that can help, or if you think you can develop such AI-based search technology, please respond to our RFI. 

We are well on our way to change the prior art search paradigm. But make no mistake: this is a long term project. Much more needs to be done, and we need all hands on deck.

When it comes to non-patent literature, much of it is in the hands of private parties. Your new database now provides a place where that prior art can be collected. In order to help our system, others who control this type of literature will hopefully contribute it to this database, or others like it.

Let me be direct: If you have literature that you might use in litigation or other disputes down the line, please contribute it to this type of database now, so that we and the public can use it up front, during examination. This will help us increase the quality and reliability of the patent grant. That’s why the USPTO applauds the efforts of MIT, Cisco, and other private companies who are helping to improve the patent system with these innovative solutions.

The Dutch Patent Office got lucky in the 1960s to find that Donald Duck cartoon, up front and during examination. Otherwise, they would have been more like the German and UK offices, and would have issued a patent that later proves to be invalid, thereby reducing the system’s reliability. Of course, if Germany or the UK had access to this new Prior Art Archive, they too might have more likely found Donald Duck’s earlier invention!

So I look forward to your collection of prior art, all in one, accessible and searchable database. And hopefully it will include much more than Donald Duck cartoons!

Thank you again for your efforts, and for the invitation to participate in this wonderful event.