The New Way to Protect Intellectual Property in the Digital Age

The transcript below has been edited for clarity.

Jim Anderton: Hello everyone and welcome to designing the future. I’m Jim Anderton, Multimedia Content Director here at engineering.com. Today, we’ll be talking about one of the most important, yet least considered aspects of engineering design: Protection of intellectual property. Joining me is Sam Baxter, Chief Technology Officer for IP.com and its sister company TextWise. With over thirty years of software and information retrieval experience including neural networks, machine learning, and deep data analysis, Sam is responsible for guiding IP.com’s development of intellectual property retrieval, vitality, and protection solutions for the business and the academic sectors.

Sam, welcome to the show.

Sam Baxter: Thanks Jim, great to be here. 

Jim Anderton: For those of us in the engineering community, we're really good at creating things, designing things, but very rarely do we actually think about how to actually protect the things that we've designed, that we've created, Patents, copyrights, trade secrets, trademarks: these are the basic things we think of when we think about intellectual property (IP). Now we're in a world where even things like code can be worth many multiples more than the actual hardware, for example, that code runs on. Do we have to change the way we think about intellectual property from an engineering perspective? 

Sam Baxter: I think it does. I think the one of the things that’s immediately surprising when we think about intellectual property is that when we're developing a solution for one thing, it might actually apply in an entirely different area of engineering and expertise. And that creates interesting opportunities both for ourselves as inventors and potentially for partners in the future. 

Jim Anderton: We're moving into a world now where code is a significant component of most engineered systems, and that's not restricted to necessarily professional engineers or even computer scientists. We're seeing people that come from other fields in other areas entering the field, especially with some of these new low code solutions. Does that change the conversation a bit when you have people who are non-professionals or from different walks of life looking at this problem?

Sam Baxter: I think there's a couple things that go on with code. I think number one. It isn't necessarily all about patents. It may be about copyrights as we saw in the case a year or so ago, I think it was Google versus Oracle, they were actually arguing over copyright intellectual property, not patent Intellectual property. 

And that's an interesting discussion, because that talks about API's and the interfaces between pieces of code, not just the way code operates with people. 

The other thing that we have to think about with code is not just the algorithms that are behind code - which certainly I have some personal experience with patenting some algorithms, or the processes behind code - but it's also about the user interfaces.

Classically in certain aspects of patenting, there's design patents. And these get exercised by people that are making things like birdhouses. Nobody really thinks much about what a birdhouse looks like, but you can get a design patent on an interesting birdhouse and this can prevent your competitor from building a different a birdhouse that looks similar. 

The same can be said for user interface components as well. So it isn't necessarily about a process or about an algorithm, but it might be about the workflow or how the interaction with the user occurs. 

Jim Anderton: There was a time when if you wanted a US patent, you actually had to fabricate a model, a physical model of the device that you're expected to patent. But it's what makes a technology or innovation patentable. Is uniqueness enough? 

Sam Baxter: So that's an interesting discussion that I think if you put a couple patent lawyers in a room, they would talk about it differently than a way a couple of software engineers talk about it. Software engineers like myself find certain aspects of patenting curious because the bar is about obviousness, it's about uniqueness and it's about having done it already. And software engineers tend to think that as long as they're given a problem, they can solve it and that kind of flies in the face of the way certain patents occur, because the way patents are granted has to do with who solved it first. 

And this was one of the big controversies, a number of years back, when Amazon got the quote one-click patent and the software industry basically said, wait a minute, what are you talking about here? It was obvious to fire off a timer that would automatically check out your card. Except that nobody done it before, and that's the difference in patent law versus in solving software problems. 

Jim Anderton: Sam, you brought up interesting point which was, for many years in the conventional mechanical world, there was a chasm between what we thought of as a first to invent state and a first to file state; that's a critical distinction, which is what you were concerned about. Should I patent if I'm the actual first to create something or can someone come after me, regardless of the legal process. Is that still a factor we still think about first to invent versus first to file? 

Sam Baxter: Well, I'm pretty sure - and I am not a patent attorney, so I don't track some of this stuff that closely - but I'm pretty sure that the world has settled on the first to invent concept. But certainly, historically you can go back and look, and I know I was taught in elementary school that Alexander Graham Bell invented the telephone. 

But I think if you actually do a little bit of research on that, you will discover that there are other people that claim they invented the telephone. He was just the first to patent it. 

So, to answer your question, I wish I felt that I had the correct answer because still may vary a bit from venue to venue. But there was the America Invents Act which basically tried to normalize the process in America with the rest of the world. 

Jim Anderton: When you think of the of code, which of course is where so much the action is today, it's traditionally, I think of code is in the way we used to do it in the days of Hollerith codes and Fortran: you get a flow chart and you conceptualize your algorithm and then at that point, maybe you wrote it in sort of a higher-level language and maybe it disassembled into machine code. But it's the algorithm was based on the flow chart based on a process of Boolean logic; A follows B follow C methodology. What part of this is actually patentable? Is the copyright of the patent based in the high-level code? Is it based in the more fundamental components of the algorithm? This is not a sort of a clear area, it seems to me, because it seems to me that for many types of programs, they're inevitably going to be similarities in the approach. 

Sam Baxter: So, it isn't about language and it isn't about a particular implementation; it is about an approach. And sometimes it's an approach or a process that combines a bunch of known approaches. This is actually true in our software. We have a patent which was granted both in the EPO and in the United States - and that's relevant and we can bookmark that and come back to it - which takes some known algorithms but combines them in a unique way and the description of how to combine them, which is expressed via a flow chart and written in a manner that anyone who understands coding should be able to replicate it. 

That is what we were able to acquire our patent for. So, to answer your question, it's not about code. You're not asked - you know, going back to your example of 100 plus years ago when they wanted a physical example of everything, which I guess you can still see some of those examples at the Smithsonian. 

Today, you're not asked to write code, and in fact I've seen very few submit code with a patent, and I've seen very few patents that actually express a lot of code per se. It's all about the flowcharts, and it's all about how the algorithm fits together and accomplishes its goal. And the goal has to be measured as being unique, a combination that someone else has not used so you're combining and creating some kind of unique thing. 

Jim Anderton: Those of us from the manufacturing side of the world where I come from, we used to say that the best way to tell our competitors what we're doing is to apply for a patent, and in some cases where we had some unique manufacturing technologies rather than patent it, we simply isolated it and prevented people from seeing it as a way to try and maintain a competitive advantage.

Some people believe that with things like blockchain technology there may be non-legal or non-patent or non-copyright ways to protect your IP without actually using the formal process. Is that going to work? Do you think that will happen? 

Sam Baxter: Well, I think they're certainly trying it right with these NFT’s, which is largely again about copyright about having a unique expression of something and saying “I own this particular expression” of it. 

So primarily I think what that accomplishes for us is establishment of a date and establishment that we did something at a particular time. And if you use certain clever methodologies like storing an encrypted version of whatever it is you're putting on the blockchain, or hashing it or something like that. It may be that you're establishing a date, but not actually exposing the details. I think this idea of trade secret is a very relevant thing to think about. Certainly, SpaceX has said “we're not patenting anything because we don't want other people to know how we're doing things and we believe if we expose it to the world, there will be certain people who don't respect our intellectual property rights and therefore the only way to protect it is to keep it as a secret”.

On the other hand, it's pretty nice that the folks that developed blockchain - whoever they may be - I know there is a name behind it, and it's never been pegged to a single person. They chose to give it to the world and there's a number of cases of that. The ZIP file format, which we're probably all familiar with was a gift to the world by Phil Katz who actually developed it. 

 Jim Anderton: Historically patents seem to be steeped in tradition – it’s a very old and established process – and lawyers seem to work with pencil and paper almost by preference. At ip.com you’re developing modern technologies that automate this process, at least from the perspective of your end users. How do you get lawyers away from doing something they’ve been doing and bring it into the 21^st century?

Sam Baxter: Yeah, so what's fascinating about the patent database and the world’s patents, is that they are expressive of ideas that may no longer be unique today, but once were very unique. And certainly, the things that are expressed in the patent database today - the current patent databases - those are very or can be quite unique. And this is an interesting proxy for the world information. Now your point that lawyers try to obfuscate this, or maybe they don't try to maybe it's their job to - I'm not going to judge that - but the way patents are written kind of falls into a couple of areas one is where the general description of an invention is presented, oftentimes with background and context as to what that invention is designed to perform. And then there's the legalese around it, what is claimed to be unique.

To many of us who practice engineering - in my case, software programming. When you read claims, they kind of look pretty funky because those are in that legalese. 

But the body of the patent is largely about the description of the problem, and that description is really rich with materials. It's rich with keywords, it’s rich with conceptual associations, and it can be mined using modern technologies like machine learning techniques and neural network techniques, to allow us to understand in a much simpler way than was done in the past with software tools as to what these patents are saying or what they're trying to accomplish for us.

It also allows us to see through the boundaries of technologies in patents; and this is interesting. If you're thinking of a system to put ink on paper, for example. That's a printer, right? And there's all kinds of interesting techniques that have been developed and patented over the years by companies like Kodak, Hewlett-Packard, Ricoh, and many others - they have developed these tiny little streams that they can precision put at a location. 

Let's imagine that we're in agriculture now, and we want to do micro irrigation and we have to put a tiny little stream of water directly on a plant. Wow, there's actually cross pollination, if you will, between those two ideas which may not be obvious to the inventors at all, and modern algorithms can help us see through this.

So this can help us understand whether what we've invented is unique or whether what we've invented is a combination of unique things making something which is now unique. And that can help us to learn how to express ourselves as well. 

One of the challenges that we hit in this world economy is that while English is the language of business, it is not the language of everyone’s thought and everyone’s expression. Many, many people in the world think and express themselves more commonly in languages other than English.

To have some method of showing those inventors how to best express themselves in that common language by using machine learning techniques and by mining databases like the patent databases holds a lot of promise for unlocking innovation out of people that are having difficulty unlocking it today. 

Jim Anderton: Sam, you brought an interesting point. First of all, you brought up globalization, of course, and the cross jurisdictional problems here. Is it possible that this software-based approach that you're pioneering at ip.com can standardize that lexicon that terminology? I remember years ago a General Motors engineer telling me that GM had to create their own internal engineering glossary just to standardize the terminology so that engineers could communicate meaningful information to each other. Is that going to be a stumbling block? Do you think for getting some kind of a universal global patent copyright system? 

Sam Baxter: So I think today that the keys to that are held by the lawyers. Because we, as the engineers, turn over our descriptions in whatever language we've written them in, and the lawyers tend to translate them and then move them into the language which is required for law. 

But it would really benefit us if we're able to better communicate both with our peers, who may again speak a different language or natively speak a different language or better communicate with the lawyers by understanding what is our local jargon and our local speak, which is what you're describing at GM - and I have worked for large computer companies over my career and I can tell you the computer folk have the very same thing where you never know what anyone is talking about until you've worked there for a while. I'm sure that's true with many big engineering-based companies. 

So, having again this worldwide patent database which has already gone through a process of normalizing it to express things in a way which is reasonably common, we can use this database to help inform inventors what is local jargon and what isn't.

For example, we work with some major auto manufacturers - and this is one of my favorite examples because I didn't know what it was. We were collaborating with them and we became aware of a submission that they worked with us on. They used the word “frunk” and I never heard of what a frunk was, and of course it's a front trunk in a car. 

But if you look in the patent databases, it is a very uncommon term and this teaches us that if we want to tell someone about a frunk, we probably should tell them about a “front trunk”, because that's more descriptive and more universal. So, the patent materials can teach us or can teach our software how to better express things, and we can build user interfaces like we are at ip.com to allow people to leverage that without actually looking at it as patents; instead looking at it as a resource for understanding phrases for understanding local terminology, and, for making suggestions as to how you may better express yourself. 

Jim Anderton: In software, one of the things it is given to areas like manufacturing engineering, it's turned things into true black boxes: throw an input in get a desired output out, whether that output are cars, reinforced concrete, whatever the need is, and code has allowed us to do that. And in my experience inventors tend to wish that the patent process was the same way, a black box: I throw my idea into the machine is at that point magical things happen and then I get legal protection out the back end. 

It's a lot more complex than that with code. With systems like the ones you're developing, are we going to get to that point where we see something where it's literally a black box? I throw my rendering into the device - into the algorithm - and then out pops my protection?

Sam Baxter: I think it's more likely that what we're going to accomplish is educating people as to what is patentable and what isn't patentable, which should result in better inventions and better patentable material. So, the process in the past was, “hey, I have a great idea. I have this three-year-old car and I'm driving it around and when it starts raining, golly I wish those windshield wipers would just come on automatically” so I think this is a great idea. 

It never occurred to me that if I had a brand-new car, it might already have that idea embedded in it; and being able to describe to a software system what it is I think I've invented and have that software system tell me now you didn't invent that. That's going to spur me to try to become more inventive, and maybe I can come up with a better way of turning on those windshield wipers. Or I can say “you know, no one's doing it on the back windshield wiper”. And I'm sorry my metaphor is breaking down, but the point is that what I think we can do is, we can improve the efficiency of the process by making it natural for engineers and inventors and creators to understand what is unique and what isn't unique and what of their ideas might be patentable and what isn't patentable.

And towards that end we've developed scoring algorithms that help people understand whether or not something is appears to be unique. And we've developed algorithms that allow that assist people in creating better invention disclosures. Because what you're really talking about there is a well understood process. I write down my idea, I hand it to the attorney, I let the attorney do all the lifting on the patent. 

But if what I really want to do is be a more innovative organization, I'm best off taking some of that work off the attorney and moving it into the engineer’s hands, so they understand when they're creating unique things and they understand how to better communicate those things both to their peers and to the attorneys so that they get better quality patents and they produce more inventions. 

You know, one of the world's largest patentors, at least in the United States, is IBM. Every single year you read that IBM is the biggest patentor. If you actually do some numbers and think about what it takes for them to produce 9000 granted patents a year, that is an enormous task. 9000 granted patents probably means they have, conservatively, they've applied for at least twice that number, so let's move it into 10s. And let's say they had to apply for 20,000 patent applications in order to produce the 9 to 10,000 that they produce. 

Now if we go back and think about the number of ideas that they had to write down and produce, well, there's probably 10 disclosures for every patent application. And that means I'm pretty sure that's 200,000 disclosures?

If I have to funnel 200,000 disclosures through a lawyer that is going to keep me slowed down. So what I really want to do is teach my inventors what is and isn't unique and what is and isn't patentable, but I don't want to get in their way. I want to make it natural for them. Part of their process of communicating their invention to their peers also becomes a process of communicating that invention for potential patentability. 

Because, let's face it, even if it isn't patentable, it still may be valuable to us just because it's in the prior artdoes not mean that it isn't something that we can utilize.

So, if I have a great idea for my particular manufacturing line, which turns out not to be patentable but is absolutely free for me to use because I discovered other people have done it, why wouldn't I bring that to my company and say, hey, look, I can improve how this manufacturing line works because we can combine this, this and this. No, we can't patent it, but yes, we can save a lot of money.

Jim Anderton: Sam, we sometimes talk about the inventor as that Lone Wolf that genius working in the basement. There's a lot of historical precedent for that. We think of Thomas Edisons of the world for example. But in reality, of course it's a team sport. It requires collaboration and multiple minds working together to create innovation. And it's that collaboration that’s easy to do in a in a large work environment, and it's harder to do if you're doing it virtually or if you're interacting with a machine, with an algorithm. 

What do you see as the change in the way that we brainstorm ideas now, if we're going to be talking about people working with AI or algorithms in general; are we going to switch this? Is this going to be engineers talking to machines? Will engineers still talk to other engineers? How will it happen? 

Sam Baxter: So Jim, I think that, like with everything, it'll become a hybrid of that lone Wolf inventor and those large group collaborations. 

The thing that the machine can add and the thing that software like ours which uses the patent databases as a proxy for unique and interesting information, they can that machine can add to the conversation. So when you're in Lone Wolf mode because sometimes you are, you can bounce ideas off of the machine and have the machine give you back things that are related to your ideas. And then after you've formed an idea, found some kind of unique combination or you figure “I think I've solved this problem” then you can use collaboration tools just like the one we're using here, which allows you to share that information amongst your peers and have them bounce it off of the machine. Then you've actually used the process from the beginning of the invention through to the idea of writing the disclosure where you're using terminology that's well understood by your peers as well as your greater peer group and you're using ideas and understanding ideas of what is and isn't unique or practice-able, what is and isn't patented as a part of your inventive process. 

So this is going to help you not just solve a problem which I think we were speaking about this idea that sometimes you solve a problem you using solutions which are not patentable and do not infringe on anybody patent that doesn't make them any less valuable, or, you solve your problem or create something new and it is patentable. And both of those outcomes are high quality outcomes in terms of us as individual inventors and as a large group of participants. 

Jim Anderton: Sam, a final question and this is. There's a size and power mismatch between size of companies like IBM or General Electric or GM and individual inventors, individual engineers or individuals that work for smaller companies. What you're talking about, the kind of technology, you’re developing, this would seem to me to have the potential to be a democratizing force that levels the playing field to a certain extent, so that smaller players may have access to patent protection at a cost and a timeline that that lets them compete with larger players. Is this a fact you think that we're going to see a democratization of the innovation process as a result of automated patent processing?

Sam Baxter: Well, I think it's fairly common that things that are done in large organizations are eventually done in small and medium sized businesses. 

And this is what you're talking about, I think, is the ability of small and medium sized businesses and perhaps even individual inventors to understand rapidly whether they have something which is patentable or not, and expressing it in a way that can reduce their costs in obtaining that patent, or a system making good decisions about that patent, or about that idea, because one way to compete with the big guys is potentially not to patent it, but to just publish the blasted things so that they can't patent it and sue you. 

And that's an interesting approach, and ip.com has been in that business for 20 years where we produce the prior art database where you can inexpensively place material into the public domain and that is a relevant intellectual property strategy just like you talked about trade secreting something you can put it out there for the world which prevents anyone else from monetizing it. And sometimes, especially in software, I think we note that people that have things that are done and working are the people that win in software. 

So maybe if you want to make sure, in software, that you're protected, what you want to be doing is expressing your algorithms in a way that you can investigate patentability, but then consider publishing those algorithms instead of patenting them to assure nobody else tries to prevent you from practicing your art. 

Jim Anderton:  It's exciting future. Sam Baxter, Chief Technology Officer at IP.com, thanks for joining me for a really interesting conversation. 

Sam Baxter: It's been a pleasure. Thanks, Jim. 

Jim Anderton: Thank you for joining us on Designing your Future. See you next time.