A Blog by Jonathan Low

 

Aug 21, 2019

Why Are There So Many Weird Tech Patents?

They are anticipatory, aspirational, defensive, offensive, but most of all a window into how companies are thinking about the future and what problems they need to solve to optimize it. JL

 Rose Eveleth reports in Slate:

Patents can be weapons and signals. They can spur innovation, as well as crush it. There is a host of strategies that lead to “zany” or “weird” patent filings (which) offers a window into how companies think about the future. Companies will patent as many ideas as they can at the early stages and then choose which ones make sense. Companies file for patents just in case. Many patents are a way to keep competitors from developing something more than a way to make sure you can develop that thing. There is value in reading the patents that companies apply for because they tell you what problems the company is trying to solve.
Amazon is putting humans in cages to protect them from machines!
Facebook is selling your face to advertisers so it can CGI you into ads!
Sony has a system where you can skip ads if you stand up and yell the brand’s name!
None of these things are technically true—they’re headlines driven by patents filed by these companies. In each case, the company has not developed these technologies (and in some cases, the headlines aren’t even accurately describing the patent, but we’ll get to that in a minute). And it’s likely that they never will. And yet, head-scratching and sometimes hilarious patents continue to populate the patent office and generate headlines. So why are there so many strange, somewhat terrifying patents that companies will likely never act on?
There are lots of reasons to patent something. The most obvious one is that you’ve come up with a brilliant invention, and you want to protect your idea so that nobody can steal it from you. But that’s just the tip of the patent strategy iceberg. It turns out there is a whole host of strategies that lead to “zany” or “weird” patent filings, and understanding them offers a window not just into the labyrinthine world of the U.S. Patent and Trademark Office and its potential failings, but also into how companies think about the future. And while it might be fun to gawk at, say, Motorola patenting a lie-detecting throat tattoo, it’s also important to see through the eye-catching headlines and to the bigger issue here: Patents can be weapons and signals. They can spur innovation, as well as crush it.
Let’s start with the anatomy of a patent. Patents have many elements—the abstract, a summary, a background section, illustrations, and a section called “claims.” It’s crucial to know that the thing that matters most in a patent isn’t the abstract, or the title, or the illustrations. It’s the claims, where the patent filer has to list all the new, innovative things that her patent does and why she in fact deserves government protection for her idea. It’s the claims that matter over everything else.
Take, for example, the Sony patent that, in 2012, generated tech headlines like “Sony Wants to Let You Skip Commercials but Only if You’ll Stand Up and Shout Brand Names at Your TV.” The patent itself is for a “System for converting television commercials into interactive networked video games” and includes a very enticing illustration of a man standing up in his living room and yelling “McDonald’s!” with arms raised, like a football referee announcing a touchdown.
But the patent’s claims are a bit more subdued: “A method executed by a processor for delivering interactive multimedia content”—in other words, the ability to insert some kind of interactive game into a piece of media content. So sure, one version of that might be yelling McDonald’s at your television. Another might be offering some kind of interactive voting for a television show contest (think: American Idol) or a children’s game to keep them entertained between commercials.
For a long time, companies didn’t really worry about the PR that patents might generate. Mostly because nobody was looking. But now, journalists are using patents as a window into a company’s psyche, and not always in a way that makes these companies look good.
So why patent something that could get you raked across the internet coals? In many cases, when a company files for a patent, it has no idea whether it’s actually going to use the invention. Often, patents are filed as early as possible in an idea’s life span. Which means that at the moment of filing, nobody really knows where a field might go or what the market might be for something. So companies will patent as many ideas as they can at the early stages and then pick and choose which ones actually make sense for their business as time goes by.
In some situations, companies file for patents to blanket the field—like dogs peeing on every bush just in case. Many patents are defensive, a way to keep your competitors from developing something more than a way to make sure you can develop that thing. Will Amazon ever make a delivery blimp? Probably not, but now none of its competitors can. (Amazon seems to be a leader in these patent oddities. Its portfolio also includes a flying warehouse, self-destructing drones, an underwater warehouse, and a drone tunnel. When I asked Lindsay Campbell, a robotics PR manager at Amazon, to comment on the company’s strategy, she said, “Like many companies, we have a variety of patents publish and the media attention they receive vary which in turns, means our PR approach is different every time.”)
David Stein, a patent attorney, says that he sees this at companies he works with. He tells me that once he was in a meeting with inventors about something they wanted to patent, and he asked one of his standard questions to help him prepare the patent: What products will this invention go into? “And they said, ‘Oh, it won’t.’ ” The team that had invented this thing had been disbanded, and the company had moved to a different solution. But they had gone far enough with the patent application that they might as well keep going, if only to use the patent in the future to keep their competitors from gaining an advantage. (It’s almost impossible to know how many patents wind up being “useful” to a company or turn up in actual products.)
As long as you have a budget for it (and patents aren’t cheap—filing for one can easily cost more than $10,000 all told), there’s an incentive for companies to amass as many as they can. Any reporter can tell you that companies love to boast about the number of patents they have, as if it’s some kind of quantitative measure of brilliance. (This makes about as much sense as boasting about how many lines of code you’ve written—it doesn’t really matter how much you’ve got, it matters if it actually works.) “The number of patents a company is filing has more to do with the patent budget than with the amount they’re actually investing in research,” says Lisa Larrimore Ouellette, a professor at Stanford Law School.
This patent arm wrestling doesn’t just provide low-hanging fruit to reporters. It also affects business dealings. Let’s say you have two companies that want to make some kind of business deal, Charles Duan, a patent expert at the R Street Institute, says. One of their key negotiation points might be patents. If two giant companies want to cut a deal that involves their patent portfolios, nobody is going to go through and analyze every one of those patents to make sure they’re actually useful or original, Duan says, since analyzing a single patent thoroughly can cost thousands of dollars in legal fees. So instead of actually figuring out who has the more valuable patents, “the [company] with more patents ends up getting more. I’m not sure there’s honestly much more to it.”
Several people I spoke with for this story described patent strategy as “an arms race” in which businesses all want to amass as many patents as they can to protect themselves and bolster their position in these negotiations. “There’s not that many companies that are willing to engage in unilateral disarmament,” says Joe Mullin, a policy analyst at the Electronic Frontier Foundation. Mullin writes EFF’s Stupid Patent of the Month column, highlighting patents that should probably have never been granted, like one for a “potent drug” that “rebukes cancer, cancer cells, and kills cancer.”
While disarmament might be unlikely, many companies have chosen not to engage in the patent warfare at all. In fact, companies often don’t patent technologies they’re most interested in. A patent necessarily lays out how your product works, information that not all companies want to divulge. “We have essentially no patents in SpaceX,” Elon Musk told Chris Anderson at Wired. “Our primary long-term competition is in China. If we published patents, it would be farcical, because the Chinese would just use them as a recipe book.”
But where do all these odd ideas even come from? Who thinks up the Amazon cage, or the lie-detecting throat tattoo? Who at IBM came up with the “drone delivery of coffee based on a cognitive state of an individual?” At some companies, teams are brought into so called “invention harvesting meetings” in which they’re asked to generate ideas or divulge inventions for the company to patent.
For patent attorneys, these meetings can be a blast. “In those sessions, the inventors are engaged and excited about it. It’s energizing to hear them talk about them,” Stein says. But one software engineer told me that not all inventors are excited to be in those meetings. “The company hired us for a specific job, to build something they have, and we have set up our salaries and our hourly rate based on that job,” she tells me. “Software engineers are not really hurting for work these days, so if you come to us and are like, ‘Make us some more money by coming up with some amazing new idea,’ I don’t really want to really give you that amazing new idea. … [M]y brain is my job and that’s what I’m lending to you. And if you want something that’s a really amazing idea, you better pay me for it, but they don’t.”
In most cases, once the inventors and engineers hand over their ideas and answer some questions, it’s the lawyer’s job to build those things out into an actual patent. And here is where a lot of the weirdness actually enters the picture, because the lawyer essentially has to get creative. “You dress up science fiction with words like ‘means for processing’ or ‘data storage device,’ ” says Mullin.
Even the actual language of the patents themselves can be misleading. It turns you actually can write fan fiction about your own invention in a patent. Patent applications can include what are called “prophetic examples,” which are descriptions of how the patent might work and how you might test it. Those prophetic examples can be as specific as you want, despite being completely fictional. Patents can legally describe a “46-year-old woman” who never existed and say that her “blood pressure is reduced within three hours” when that never actually happened. The only rule about prophetic examples is that they cannot be written in the past tense. Which means that when you’re reading a patent, the examples written in the present tense could be real or completely made up. There’s no way to know.
If this sounds confusing, it is, and not just to journalists trying to wade through these documents. Ouellette, who published a paper in Science about this problem recently, admitted that even she wouldn’t necessarily be able to tell whether experiments described in a patent had actually been conducted.
Some people might argue that these kinds of speculative patents are harmless fun, the result of a Kafkaesque kaleidoscope of capitalism, competition, and bureaucracy. But it’s worth thinking about how they can be misused, says Mullin. Companies that are issued vague patents can go after smaller entities and try to extract money from them. “It’s like beating your competitor over the head with a piece of science fiction you wrote,” he says.
Plus, everyday people can be misled about just how much to trust a company based on its patents. One study found that out of 100 patents cited in scientific articles or books that used only prophetic examples (in other words, had no actual data or evidence in them), 99 were inaccurately described as having been based on real data. For example, in a 2014 paper in the Journal of Chemical Sciences, researchers wrote that “Dehydration reaction in gas phase has been carried out over solid acid catalysts,” citing a patent for proof. But that patent’s example was prophetic, not real. And Ouellette points out that the Theranos patents were full of prophetic examples like “Micro-needles automatically draw small quantities of blood painlessly. A mechanical actuator inserts and withdraws the needle.” She says that if she went into reading the patent with no context, “[E]ven as an expert, I wouldn’t be entirely sure if they had any technology working or not,” she says, which might have led investors, reporters, and the public to think that Theranos had actual, working technology. (It did not.)
Which brings us to the media coverage of these patents. Not only do most stories about flashy, strange patents misrepresent what filing a patent might actually mean, they also often misrepresent what the patent is actually for. “Sometimes I look at patents and think, what is going on with this internet outrage?” says Stephen Yelderman, a law professor at Notre Dame who used to work as a patent attorney in Silicon Valley. “It seems like people aren’t reading the patent and giving its fair due.” Patents are an odd type of document, and it takes some practice to understand what to look for. Remember, it’s just the claims that matter, but news stories often home in on other parts. “There will be an article written based on the title or the abstract, saying that this patent is claiming a broad idea that seems to cover half the internet, but when you look at the claims, they’re very specific,” says Ouellette.
Other times, coverage focuses on the illustrations. “It is very important to note that the pictures in a patent often don’t give you a great sense of what the patent actually covers,” Ouellette says. Often the pictures show one thing, and the claims the patent actually makes are a bit more tepid. Take the Amazon worker cage patent. The image is striking, a person standing inside a metal cage. But when you read the claims that Amazon is actually making in the patent, there’s nothing that specifically describes a cage at all. It turns out, those images are rarely ever created by the inventors themselves but are instead drawn up by the patent attorney or a special patent illustrator. (Facebook representatives did not respond to our requests for comment. A Sony representative told me that “we cannot comment on the details of these matter due to the confidentiality.”) Yelderman also says that sometimes patent attorneys have a bit of fun with the illustrations. “I once made an image that … for no reason, happened to include the text of the Foreign Corrupt Practices Act, just to see if anybody would notice, and it had nothing to do with the invention.” As far as he knows, no one has noticed.
Stein says that recently he’s had companies bail on patents because they might be perceived as creepy. In fact, in one case, Stein says that the company even filed to have the requests pulled offline by the patent office. As distrust of technology corporations mounts, the way we read patents has changed. “Everybody involved in the patent process is a technologist. … We don’t tend to step back and think, this could be perceived as something else by people who don’t trust us.” But people are increasingly unwilling to give massive tech companies the benefit of the doubt. This is why Google’s patent for a “Gaze tracking system” got pushback—do you really want Google to know exactly what you look at and for how long?
Let’s go back to the infamous “Amazon wants to put workers in a cage” patent. Looking at the claims of the patent, you’ll see a document that describes essentially a moving platform that can transport a worker around an active work site, which might include dangerous machinery. The company filed for the patent in 2016, but it didn’t make headlines until two years later, when it was highlighted in a presentation by A.I. researchers. They framed the patent as “an extraordinary illustration of worker alienation, a stark moment in the relationship between humans and machines.” When the press ran with that interpretation, Amazon’s senior vice president of operations Dave Clark made a statement on Twitter: “Sometimes even bad ideas get submitted for patents. This was never used and we have no plans for usage. We developed a far better solution which is a small vest associates can wear that cause all robotic drive units in their proximity to stop moving.”
It’s a good example for why there is still real value in reading the patents that companies apply for—not because doing so will necessarily tell you what they’re actually going to make, but because they tell you what problems the company is trying to solve. “They’re indicative of what’s on the engineer’s mind,” says Duan. “They’re not going to make the cage, but it does tell you that they’re worried about worker safety.” Spotify probably won’t make its automatic parking finder, so you don’t have to pause your music in a parking garage while you hunt for a spot. But it does want to figure out how to reduce interruptions in your music consumption. So go forth and read patents. Just remember that they’re often equal parts real invention and sci-fi.

0 comments:

Post a Comment