Many hands make light work

They say that too many cooks spoil the broth. Too many people, regardless of expertise, don’t always produce the best end product. Patent owners know this, and it leads some to pick a small team of specialist researchers when enquiring about a patent’s validity. Essentially, they trust a select few.

But if too many cooks spoil the broth, many hands might also make light work. Article One Partners (AOP), a company that coordinates worldwide research on prior art, prefers a crowdsourcing model.

Using 22,000 researchers, hailing from 180 countries, the company aims to match clients’ requests by spreading its wings far and wide. The world is its oyster.

Placing their trust in this expansive network, patent owners ask for help. Whether they are launching a new product, trying to invalidate a rival’s patent, or defending a patent in court, companies need information quickly.

“Generally, studies take four to six weeks,” says Cheryl Milone, who founded AOP in 2008 to add a “crucial level” of review to the US patent system—strengthening “legitimate” patents and reducing “unjust” patent monopolies.

“In my years as a patent attorney, I spent countless hours and late nights searching for prior art. I was struck by how deficient the process was, and knew there had to be a better way. The combination of globalisation, technology and the economy has paved the way for new and disruptive business models such as crowd sourcing to thrive,” she says.

Searching for prior art can be like finding a needle in a haystack. But aided by thousands of researchers, who are motivated by financial reward, patent owners can be confident the crowd will find a smoking gun, if one exists.

Milone refers to the research community’s “different competencies”, the first being that the more seasoned researchers “know most of the standard places to look, which are the core types of references from a traditional search”.

“Other members of the community, including some of those traditional researchers, will use proprietary databases or hard-to-access databases, maybe through a library, to give themselves an advantage of finding parts and materials.” The ideal scenario, she says, is for someone to read one of the requests and say, “I invented that”.

“When you aggregate all of these sources and extend them to a worldwide level, you’re able to dramatically increase the quality and the results.” AOP, which has rewarded researchers with around $2.8 million to date, demands they send in references from all their searches.

Many will examine existing US patents—which are all fully digitised—so it’s perhaps unsurprising that the most valuable references come from non-patent literature, which is typically harder to find. “For non-patent literature, foreign language articles are generally translated into English in abstract form only, so the full text and footnotes are missed,” Milone says.

“There is a tremendous amount in foreign language publications that isn’t digitised and, even for English language articles, figures and tables aren’t digitised. So, really, there is a treasure chest of information which can help people understand the value of patents.”

When it’s received, a reference goes through an automated filter at AOP to see how closely it fits the technology description, before it’s sent back to the client.

The process typically takes four to six weeks as, for example, requesting hard copies from libraries can take two weeks. Afterwards, the clients can go back and ask the researchers for more articles by the same author, or for more information, so there is the ability to “iterate” the results, says Milone.

“Having the four to six-week timeframe allows more of that iteration, which can provide additional value to clients. But occasionally, if a client has a more urgent need, we can cut the wait to as little as a few days.”

"Because of the aggregation of so many resources, even where a smoking gun is not found, the client has more confidence that every stone has been turned.”

Pointing towards the benefits of crowd sourcing, she says it enables researchers to unearth information that might otherwise be tricky to find. This is particularly true in less-developed regions where there may be inadequate databases or non-digitised information. “In these places, crowd sourcing is beneficial because the native language speakers go beyond those limitations and are able to look at hard copies,” she says.

Milone uses the examples of Russia, the Middle East, South America and Africa, and even “developed regions for prior art such as Japan, China and Korea”, where she says the community can produce better results than a firm based in those countries. “Again, it’s the aggregation of resources.”

Some would argue that smaller teams of researchers are more effective than a crowd. But while Milone admits that small teams conduct high quality research, she believes they are limited to which databases they can use. “It’s not the researchers, it’s the limitations placed on them— and that’s what crowd sourcing overcomes.”

Most importantly, she says, AOP’s work allows patent owners to fully understand their objectives. “For example, in validity it was difficult historically to determine whether the search was accurate unless you found a smoking gun. Now, with this platform, because of the aggregation of so many resources, even where a smoking gun is not found, the client has more confidence that every stone has been turned.”

Broadcasting is another benefit of the crowdsourcing model: AOP’s platform amounts, in effect, to a public notice that patents are being researched. In turn, for companies routinely targeted by aggressive, non-practising entities (NPEs), it becomes easier to highlight the risks associated with an NPE’s actions. “So it’s helping to defend and deter those actions,” says Milone.

When cases come to court, how direct is the relationship between AOP’s research and patents being invalidated? “It’s hard to judge,” she says. “It’s not a one-to-one relationship. Litigation is a long process.”

While this may be true, Milone says she does know of patents that have been invalidated with the help of evidence provided by AOP. And she says there can be a more direct relationship after the America Invents Act, with its ‘first inventor to file’ system, is implemented in September 2012.

“It has very clear proceedings for adjudicating patents, and those will be helpful because they will be having a close look at patent validity and making a decision to allow the stakeholders to move on and not waste resources on patent disputes,” she says.

“Courts are seeing assertive claims over patents, and it is exciting to work for a company that is helping to evaluate these innovations (while they’re defending their patents) and to build their portfolios to support innovation,” she adds.

“AOP is a fundamental component of building patent quality, so if there are higher quality patents and fewer lawyers in the system, then the US Patent and Trademark Office (USPTO) can return to its original mandate, which is to drive innovation rather than driving litigation and being driven by litigation.”

While AOP has made positive strides since forming four years ago, it is still a young company looking up to bigger players in the market. Its presence in Europe is low key, and although it operates mainly in the US, there are still other, more established rivals working there.

The birth of crowd sourcing

According to Roya Ghafele, an academic at Oxford University who has worked with AOP, it was Professor Beth Noveck who first wrote about crowd sourcing for prior art searches. Noveck, formerly head of the Open Government Initiative under President Obama’s administration, published an article in the Harvard Journal of Law & Technology in 2006.

In it she described a “crisis of patent quality” at the USPTO, worsened by restricted access to information. “Examiners may neither consult the public, talk to experts nor, in many cases, even use the Internet,” she says. In Noveck’s view, the answer was for the office to adopt a “peer to patent” model: crowd sourcing.

In the same vein as Noveck and AOP, Ghafele says crowd sourcing, which she compares to online reference source Wikipedia, is the “right thing to do”. She says that, potentially, it can help patent offices in developing countries to better examine their patents, contributing to higher patent quality and, ultimately, to bringing down the cost of litigation.

“This game would be reduced if the patent offices issued patents of better quality,” she says.

But, for AOP, here lies something of a paradox. While the company helps to find information that patent offices have missed, it also relies on their missing this information to expose their mistakes. Of course, patent offices can’t be expected to find all the prior art in the world, and there will always be a market for filling in these gaps. But the more efficient the patent office is—to some extent at least—the less AOP and its peers will be required.