Speaker 1 0:00 Good afternoon and welcome to today's spotlight session webinar, non patented software innovations how to protect and extract value presented by Autumn. My name is Sammy Spiegel, autumns Professional Development Manager and I'll be your staff host for today. We're continuing the great discussion started during our annual meeting in New Orleans with a new online opportunity autumn spotlight sessions. During today's session, you will watch a popular panel recorded at the annual meeting, and then participate live with our panelists for a dynamic question and answer session. All lines have been muted to ensure high quality audio and today's session is being recorded. If you have a question for the panelists, we encourage you to use the q&a feature on your zoom toolbar or the raise hand function to ask your questions aloud. If you have a technical question or comment, please feel free to use the chat. Should you need closed captioning during today's session zoom closed captioning feature is turned on and available on your toolbar. Before we begin, I would like to take a moment to acknowledge and thank autumns online professional development sponsor Marshall Gerstein IP, we appreciate your ongoing support. I now have the pleasure of introducing you to our spotlight session presenters who are joining us for today's q&a. George malapa leads new videos research licensing efforts. In this role, George works with his new video colleagues to negotiate and license research projects. In addition to his role at Aveda George is an active member of the Technology Transfer community and mentors Early Career Technology Transfer professionals through the autumn technology transfer Fellows Program. In addition, George serves as the co chair for the autumn Metta World project, and serves as an autumn Foundation Board member. Laura Dorsey is Director of Innovation development software and copyright at ko motion. Prior to joining ko motion, Laura ran a grant program that funded the university side of the industry University projects at Washington State Universities and worked at Honeywell. Laura manages a portfolio of various technologies ranging from computer engineering, healthcare IT software and digital content to cardiology and oceanography. She also provides valuable resources for stakeholders and achieving innovation milestones, including co creation of the idea to plan curriculum, have their back to a partner at Quarles and Brady LLP is Phoenix office chair and the Intellectual Property Group A member of the data and privacy and security industry team. She is a transactional practice covering the areas of E commerce, software technology and data privacy, with an additional focus on copyright and trademark prosecution. More specifically, Her work focuses on software agreements IT outsourcing, internet terms and conditions, internet privacy policies, electronic contracting technology development agreements, copyright licenses, copyright, prosecution, trademark licenses and trademark prosecution. Welcome all, we're so excited to have you joining us again today and appreciate you taking the time to come back for a live q&a session. I will now launch the presentation delivered at the annual meeting in New Orleans attendees. As a reminder, please feel free to submit your questions using the q&a feature on your toolbar for our panelists throughout the presentation. And we will answer those at the end. Additionally, once the recording concludes, you'll be able to join live and use the raise hand function to ask your question aloud when our panelists come in live. So with that, I will get us started with this new screenshare and get started. Speaker 2 3:40 So to kick things off, why organize the session Why think about non patented software. So personally, in my own experience, I have seen more and more disclosures coming into our Tech Trends office on software innovations that either non patentable or occasionally these patentable technologies. But the lifecycle of the technology doesn't really lend itself a patent. Or it could be a mobile phone applications. All the code has already been released open source. Another area that I'm increasingly seeing is machine learning models, the models, the algorithms themselves are not unique, but they have been trained on a unique data set. So how do we think about commercializing these trained models? So pretty impressive background we have today, I think we'll hope to answer some of those questions. So what will we be covering? So we'll try and cover what types of intellectual property protection exists. Some of the necessary diligence that we need to be thinking about how do we value these software innovations as we license them And then at the end, we'll hopefully have a robust, open discussion. So to kick things off, I like to invite heathered to start. Speaker 3 5:12 Before we get started too much, there's one chair up here. So whoever can run up to the front the fastest, you can have it if you guys can't hear us if you if we have some Thank you. Let us know shout out. We'll also mention if if you have some brief questions, do feel free to raise your hand, we can try to address them. As we're having the conversation, we've obviously got a packed room today, we're not going to be able to get to everything, we'll try to save some time towards the end. But Do feel free we want we want you to get as much out of this session as you need. So if there's something that strikes your fancy as we're going along, do feel free to pipe up we reserve the right to cut the mic though, if we get waylaid. So, as we mentioned, we're going to talk a little bit about some of the basic intellectual property rights that exist. It just sort of above and beyond what you guys would typically see. So generally speaking, you're going to have four types of intellectual property rights, you're going to hear people talk, you know, maybe about some of the others moral rights, industrial design rights. Generally speaking, those are like the cousins to core IP rights. When you're talking about IP rights, you really got the four categories patents, trademarks, trade secrets, and copyrights. Patents, we're not going to touch on the whole point of this session is that we can't get there or we don't want to get there. So we're going to push those off to the side. Trademarks, you have come across those obviously, in your lives as a consumer, any brand name, logo, device, color, smell, sound, the indicates an origin, right that you as a consumer, when you see that you know what that is, if you see the color brown in the context of a box, you know, where it came, you know, came shipped via UPS, if you see the golden arches, you know the quality of the or lack thereof, depending on your opinion of the product that you're about to go in and buy, you know what proximately, what it's going to cost, you know what it's going to taste like trademarks are actually a consumer protection issue, relevant in the context of some of the things that we'll talk about, depending on what your goals are for licensing some software, and it's party time. That was not planned. But so we're not going to talk too much about trademarks, again, a relevant consideration as you're looking at maybe some of your goals with respect to software product, but not the main focus of our conversation today. Trade secrets, any information basically, that derives independent economic value from not being known, and is subject to measures to maintain that confidentiality, we're going to touch on them. Not very much, for reasons we'll talk about. But trade secrets or that third category, the fourth category, and the one that is primarily going to be the topic of our conversation today are copyrights, copyrights, any work of authorship, architecture, software, operas, you name it, vessel whole designs, shockingly, are subject to copyright protection, any original work of authorship that has some originality to it, that is fixed at a point in time. So as you're waiting in your word processor, and your file disappears, which has happened to all of us, you do not have copyrights in there, because it's no longer fixed once you hit save on that software on that file, and you have that file, and you're going to have copyrights in it. So one of the things that I wanted to make sure I pointed out is, as we look at copyrights, it's actually the same constitutional basis as patents, that Article One, Section Eight, in terms of promoting the progress of science and useful arts. So same concept, in terms of why we have them, just a different law, just a different application different, just works differently. So with that, anything. Anything you want to understand. So I always get this question when I have this conversation, right? You didn't mention data? Where does data fit into all of this? The practical answer is that it doesn't data is generally not an intellectual property, right? It is a fact it is a fact that I walked across the street at a certain point in time in the day and at a certain speed. Nobody can own that it just is you just have that the fact just exist. And as a result data itself is not protectable as an IP right. Now, that's not to say that you don't have a valuable data set, which is different from from just a raw data point. And a data set is something that you could potentially look at licensing there is some value in that you did You've spent the time to create the data that is potentially a value to someone that may want to license that for training their AI or some other purpose. And so you can look at licensing that data set. But that is also distinguishable from a database. If you have data that you have selected, arranged in a creative way, that database can have copyright protection. It's not always robust copyright protection. But there is copyright protection that can exist in a database. If you want a little light reading for your weekend, the Copyright Office has recently updated recently, within the last couple of years, updated their Compendium they have also released new guidance on particularly automated databases. And it can be a complex field. But the point is that you can have copyrights in those in those databases. So that's kind of how data fits in this paradigm of are for intellectual property rights. raw data, probably not protectable datasets, potentially valuable can still be licensed. Depending on how that dataset is selected and arranged. You could have a database, which is copyrightable, you could click on it, you may make it quick. Speaker 4 11:12 I would suggest that all of you when somebody particularly a funder uses the term data, you look at the definition for what they mean by data. Heather is describing data as kind of, you know, numbers, you know, things like that. But what about an MRI that's been annotated, right? Does your funder consider that data? And have they swept that into having some broader rights to it? Certainly, the Department of Defense will define it broadly and take those rights, which is pretty unavoidable. But when people say date, I feel like it's saying Open Source, you need to really understand what do they mean when they use that word. So Laura, Speaker 2 11:49 and Heather, you bring up a good point about where this data comes from. So Heather, we have faculty who might say, George, I've trained an AI model, I got the all the data by sweeping through Twitter. It's, would it be and sometimes they would use all its values and just using it for academic use? How should they take turns off his be looking at these kinds of statements? Should we be concerned? Are they violating the terms of use of Twitter, you Speaker 3 12:21 should be highly skeptical of any statements that you get in that regard. I, you, lawyers are never supposed to say never or always. But 99.9% of the time, if you're pulling something off the internet, if you're scraping data, if you're downloading anything, there are either publicity or privacy rights issues, or copyright issues that are going to be involved. Anything on the internet is going to have some sort of terms and conditions with it, again, 99.9% of the time. So you should always be highly skeptical of those of those of those comments. And I know, in particular, and we'll go ahead and mention it. Now. George and I were talking about the ever album case, which came just came up through the FTC originated around 2017. I think they settled it out in 2020. Ever album had one of those tools where you could upload your media photographs, videos, they would tag you can tag people. And they were using that data to train their facial recognition algorithm come to find out according to allegations from the FTC, they were scraping lots of data from lots of places and applying facial rec to it. And again, using all of that data to train their AI in violation of a number of people's privacy rights, running facial rec techniques, using facial recognition technology, super sensitive area right now with respect to is essentially biometric data that can be used to identify people causes all kinds of concerns from from a privacy perspective. So the FTC made allegations that they did all of this without providing proper disclosures and actually made some misrepresentations. As part of the settlement, not only did the FTC require them to delete the data, they actually made him destroy the algorithms that they had developed using the data, sort of fruit of the poison tree kind of argument, that if you did all of this in ways that you shouldn't have, you don't get the benefit of the algorithms that you developed. So very important, as you hear some of those. Here's some of those statements about data that went into something. Be asking those questions about where about where it came from, because you don't want to end up at you know, having something that you can't use or can't commercialize, because you've run afoul of of rights along the way. Speaker 2 14:44 And this is something we've done at Northwestern that. If you haven't already, you should update your disclosures, especially your software disclosures to understand whether any data was used and where the data came from. You know, often academics share their data, email it to someone in another institution. So you want to understand where did the data come from? Did they have rights to use it? Laura, when you do license data, you do it at University of Washington, do you use things like Creative Commons license to license these data sets? Speaker 4 15:22 What do you mean by data? So, our patient derived data, you know, data coming from our School of Medicine, our School of Medicine has asserted that that's theirs to deal with, because it's not intellectual property, right. And so they are taking a position a philosophical position that they're going to do what's right for the patients. And so they have made some data available, kind of behind the firewall to some big tech companies, so they can run their algorithms against it without releasing it. And there's this complicated governance structure they've set up, you know, so there are other cases where we have licensed data along with some other kinds of assets. And really, I think we're relying on some contract language to protect us rather than the typical grant language. And I would just add it. Second comment, my AI researchers in our computer science school, they are signing incoming data use agreements, some of which have reached through into IP creation. So make sure you know, you know, as George says, the source of the data and what conditions go with it, I feel like we're really in an evolving time, where everybody's trying to figure out how to monetize data. My AI researchers just need the data, right? They're creating, you know, great new ways to use artificial intelligence, they just need other people's data. So really watch for that, too. Speaker 3 17:08 So we mentioned licensing trade secrets, that's about all I'm gonna say about it. There's actually a roundtable tomorrow. I understand I have not verified this on the schedule. But there's a roundtable tomorrow to talk about it. It's just difficult, right and institutions that are subject to FOIA requests. You're in a, you're in academia, where it's a publish or perish kind of, kind of situation. For many, it's just incredibly difficult, particularly got transient students coming in and out to maintain something with a level of trade secret, or with the level of confidentiality that you need for a trade secret is just difficult. So we're going to leave it at that. And I guess, if you're interested, you can contribute to the roundtable tomorrow and see if you have any, any better ideas. So that kind of takes us to the meat and potatoes of what we're going to talk about just show of hands, who has a working knowledge of basic copyright law. Oh, good, then we'll skip through the basics. So as we look at, you know, it is that bundle of rights, right, you can separate it out, you can figure out, I often describe it to my clients, as a snowball, you can slice and dice that snowball in infinite number of ways, as long as as long as you still have what you need at the end of the day. And your licensee still has, you know, gets what they need. At the end of the day, you can end as long as nothing, you're not granting, you know, exclusive rights to one person and then end up needing to license some something to somewhere else, then you can do that. And you have that you have that you have that flexibility. So we're gonna skip over some of those basics. How do we get a copyright? I get that question a lot, right? How do I if I'm going to have to rely on copyrights? What do I have to do? Because most folks are obviously used to having to go through the patent process. The basic answer is you don't have to do anything. Those copyrights immediately sprang into being the moment you know, the moment you put pen to paper or save that, save, save that file. The thing that becomes more complicated, particularly in your guyses settings, is in whom do those rights best? And that's where you can get some tricky conversations. If you start with the base layer of what copyright law says, you know, going back to the Constitution, Article, one section, eight vests in the author vest in the person that's actually creating the work. And then you run into the arguments of well, what does it mean to create the work if I'm just an advisor, providing input? I shouldn't I be identified as as an author and depending on the facts, the answer maybe No, right? Because you're not actually contributing that original, that original fixed, fixed work. So taking into account kind of where those were rights actually vest is going to be critical. When you look at when you look at whether and how you need to do something else to protect these rights, you certainly can go through the registration process, show of hands, anybody gone through the copyright registration process, couple folks, copyright office, luckily, I'm fairly confident there's no copyright office representatives here, their website is a hot mess. It's just not very user friendly. There are some benefits to going through that copyright registration process. You have those copyrights immediately. But if you actually want to enforce those, you need that registration to get into federal court. There's benefits to doing that early and getting that registration early. You get additional rights, you get additional remedies, if you actually do have to have to pursue something in court, it's cheaper to do it early. But of course, you know, the, the downside is that you have to submit copies of what you're registering. And it becomes part of the Library of Congress, there are processes to help protect some of the trade secrets that may exist within your within your code, and you don't have to submit all of your code. Obviously, that would be a lot of data. But but you're still disclosing, you know, part of part of your part of your work and making that making that part of the public resources that are out there. So those are kind of the pluses and minuses. At the end of the day, whether or not you want to register that software depends on weather and how you're going to want to weather and how you're to license it. And what that looks like in terms of your practical, sort of the practicalities of how that ends up playing out. So it's something that you consider at that time, Speaker 2 21:46 just to add on the author's creators. I have had situations where faculty assume they are creators, in fact, because they got the funding. So sometimes, they have been surprised that when we do the valuation and say you're not actually one of the creators of the software. So, Laura, when we think about software licensing, that's not patented. What's the philosophy at the University of Washington on protecting commercializing? You know, pushing out these software innovations? And I actually have a part two question on that is that University of Washington is, is known for its computer science department, it has been leading for decades. But there are the universities where the focus is on therapeutics. Maybe they've had a blockbuster drug that's brought in millions, if not billions in licensing revenue over the years. So how do you convince university leadership, your tech transfer director that you go out and spend time thinking about these non patented software that probably will never bring the same kind of money? So how do you, you know, convince them that you should be spending time thinking about these innovations. Speaker 4 23:05 So at the University of Washington, the way I found my way into the job there was I was working at an economic development nonprofit, we had a building on campus. And this Maverick little group from the tech transfer office wanted space in our building, and my boss at the time, gave them space. And that Maverick Maverick he little group was a software copyright group. So I got to know this offshoot. And I think at the time, they enjoyed being an offshoot, you know, they were remote from the main office, and they were really kind of forging the way this was in the 90s. And so the University of Washington for a long time has recognized the unique value that copyright and software in particular can bring to the university about why don't know, 14 years ago, maybe the Software Group was kind of absorbed into what we might loosely have called the patent group. In both groups did some of the other work, but We're centrally you know, focused more on copyrights or patents. And the way we had been managing software we tried to help the patent side think more broadly about so the University of Washington for a long time, we have recognized the value of software and, and sometimes it can be quite valuable. I had software I just finished licensing that we started licensing in 1997. It was created by our IT department to play a role in tape backup systems, so was being used to backup our own data. And over the years, it brought in more than $16 million. So in our world, that's not small money. It is over a lot of years, but it's not small money. So our philosophy is software is an important asset. It can be monetized and so we take it very seriously However, as George points out, there are other kinds of software and copyright protected assets that are never going to bring in big dollars. And I'm happy to say in my office, we are not primarily driven by revenue, we do have the flexibility to be able to have a mixed portfolio. And as individual portfolio managers, we can make decisions about what we want to spend our time on. And so we have other kinds of assets that we think have important impact that we will spend time on, that are never going to generate enough money to really even pay for the time we have invested in them and trying to help them. And in our system, that's okay. You know, the winners don't like to hear this, but the financial winners subsidize the financial losers in our system. And that means we think we have broader impact across the research enterprise. And Speaker 2 25:58 I know that I do you notice that you have also apps like the suicide prevention app that could you elaborate on Speaker 4 26:07 right, we have suicide prevention is training. So it's not software, it's more, you know, other copyright materials. That obviously, we think is very important societally, because it trains people to look for suicide signals and be able to address those, and it has brought in some money. But again, I would guess, based on the amount of time my colleague who manages that spent on it, probably not as much money as we've spent in her salary. Unknown Speaker 26:34 Thank you. Speaker 3 26:39 So anytime you have one of these assets come across your desk. And you've figured out that you own it, and you've gone through it, and we'll talk about some of the diligence questions that come up later. The question is, how am I going to license this thing? Right? Like, what? How are we actually going to do this? The first question I would start with, and I'd be interested in your thoughts, if you agree with this, what are our goals with this asset? Right? Like, what do we do we want to try to monetize it? Do we just want to get it out there, like the suicide, you know, common to try to locate those issues? Is it more of a brand recognition thing where we just want to get out in a new field? I mean, what are some of the things in any other? Those are the ones I could come up with? Yeah, I'd Speaker 4 27:22 say we're really very driven by the researchers goals, right? We just tag along behind them using our IP bag of tricks to try to help them achieve their goals. Yep. Speaker 3 27:31 And then what concerns do we have right, like, what is, is is it a tool that's going to generate potentially a lot of liability? Is it is it something that, you know, once the code gets out, we think somebody's going to infringe it or steal it and use and misuse it in a way that we don't intend it to be used. So thinking through those goals, thinking through some of those risks, will help figure out how you actually want to go about licensing it. So we'll talk about a couple of different licensing models. The one that I immediately think about as outhouse counsel, is the proprietary licensing model. When I say proprietary, I think, you know, it's a custom license agreement, I'm only distributing the object code. It's just a traditional software license, somebody's going to get a copy of the object code, and they're going to install it on their own system, and they're going to take it and run with it. There's a couple of different ways to do that, obviously, the old school way. Traditional paper licenses where people actually use to sign with pen. And that's how and that's how you got your your agreement. This day and age, a lot of the stuff that I see is either a click rap or a browse rap. And those do have different meanings. Click raps, your signature is any affirmative act that somebody is taking to show agreement to that right, they're clicking the checkbox, they're clicking, I agree, they're clicking Submit, however your workflow is set up. It's whatever that action is that they're taking. Usually, it's a click, hence the name that actually shows what that agreement, you know, actually indicates that agreement, browse wrap is more akin to a traditional shrink wrap from back in the day when we used to buy our software Best Buy. And we would actually like take the shrink wrap off the box, you're deemed to have agreed to it just by browsing to the website, just by downloading the software. It's just more of an implied kind of kind of kind of signature process. One of the things that I wanted to point out in this conversation, those so those are the signature processes, right what in click wrap browser up, shrink wrap if you go that route? If we have a traditional paper signature, how do we prove we have a contract? Not a trick question. You go to your file cabinet, right and you pull out the contracts. And you're like Mrs. Smith, is this your signature? We can't let those good contracting processes fall by the wayside. If we're in an online setting. If you Have a click wrap, who clicked it? What IP address? Was it? What date and time? What version of the click wrap Did they agree to, you've got to collect all of that data in your electronic filing cabinet. And you've got to store it in a way where that data is going, the integrity and the accuracy of that data is not going to be challenged, right. That's how that's the only way you can prove an electronic contract in place. Same with a browser app, it just gets a little more complicated. Speaker 2 30:25 So Heather, you bring up a great point about click through licenses where people have clicked, they've downloaded the software made me may have given their name, you have collected the IP address. Should the university Tech Trends office be concerned about the data that they're collecting? Is there a privacy concern that they should be thinking about? Speaker 3 30:49 Issues is probably a better word than concerns. Anytime you're collecting data in an online setting, you need to pay attention to what data you're collecting. And what is it clear what data you're collecting, right? Go back to our ever album case, right? They got lots of data that they shouldn't have had, and they were using it in a way that wasn't disclosed to consumers. Generally speaking, data privacy issues in the US are a matter of most generally, Section Five of the FTC act. It's an unfair and deceptive trade practice. Obviously, Section Five of the FTC Act applies differently if your nonprofit for profit. And so the law itself may not technically apply. But data privacy is such a huge issue these days. It's something to be aware of, if you're if you're collecting that data. Theoretically speaking, you should always have a privacy policy that discloses what you're collecting, how you're using it and who you're sharing it with. There are more bells and whistles on the data privacy side of the world. But generally, if your privacy policy does those three things, you're 85% of the way there. But yes, it's absolutely something that you should be paying attention to. Anytime you're anytime you're collecting data online. Speaker 2 32:01 Thank you add a follow up question to that. So let's say the privacy policy does address those. are we obligated? So if if I if you are coming to the Northwestern University's website to download software, how much effort should I make it make to make you aware that we have this privacy policy? Is it enough to have a somewhere on the webpage or you know, how actively should I be promoting that to the download, Speaker 3 32:31 the general rule is the is the old school has to be has to has to be conspicuous, clear and conspicuous. I was forgetting the other word, it has to be clear and conspicuous. For the most part, that means what most people do is a link at the bottom of the website that says privacy policy, arguably, you know, argue whether that's clear and conspicuous, but that's where people expect to see that link to a privacy notice. And so that's where it needs to be, you know, the same rules for anything that has to be clear and conspicuous. It has to be the same size font is everything that surrounds it, you can't have dark blue letters on a black background, anything that you know, in anything that that doesn't draw attention to it or tries to hide it is not going to look good. So you just you want it you want it to be there, you want it to be noticeable, but you don't have to go necessarily out of your way to have you know, huge splash pages and and pop ups along the way. Thank you. So a couple points here, you will hear the terms and User License Agreement Terms of Service Terms and Conditions Terms of Use. That nomenclature from my perspective is all interchangeable. It does not. And to that end, if you have somebody who is adamant that something has to go out as an end user license agreement, fine, call it whatever you need to call it. The bigger issue is how are you getting your agreement? How are you tracking that data? How are you? How are you proving that contract exists? I have my own interpretation of those terms about what I think makes sense in what context. So in my practice, I use those those terms differently. But for the most part across the industry, like there was no legal definition to the terms of service versus the Terms of Use versus an end user license agreement. One of the things that I know Laura, you had brought up as we were preparing for this, what a software license is not is a confidentiality agreement. Confidentiality Agreement is a confidentiality agreement, software licenses, a software license, those two things are not interchangeable. Although I suppose you could just change the titles. But substantively, speaking as a general non disclosure agreement, is not going to be sufficient as a software license. It's not going to address the copyright restrictions, the copyright licenses that you need in that context. On the flip side of that a lot of a lot of the End User License Agreements that I see on a software from a software perspective may not even Have confidentiality obligations, because you're posting you know, you're making it for available for download on your, on your website, the con, you know, the fact that the software exists is not confidential. So you may not have a confidentiality agreement in that, or confidentiality provision in that agreement. So that's something to bear in mind, as you're looking at the documentation you have available to you. In some of these initial conversations, those two things are probably not interchangeable on any given basis. Speaker 2 35:30 So when you do software licenses, one of the things I often hear from potential licensees is, you've given me this bundle of rights, and you've given me the right to create derivative works. You've given me a piece of code in C, I'm going to rewrite it in Python. Now why should I pay you a license fee? Why should I continue to pay a royalty? Or if I've changed the code substantially over the next two years? Why should they continue to pay you a royalty? How do you argue against because you want the code? Speaker 4 36:02 No. So it's a good point. And I would say it's fairly common at the the big universities that license lots of software like the University of Washington, that that's a reasonable position for a licensee to take. It's also a reasonable position to argue that so we give it to you in MATLAB and you immediately rewrite it in Python, because MATLAB is not a commercial language, you know, should you owe us no royalties after that point? No, I don't think that's how this is gonna work. Right? So. So the way many universities do that is, the licensee can own the derivatives, you know, separate from the original work, but they continue to pay royalties on the derivatives as well as on the original code. And I think many universities, including the University of Washington, would renegotiate a deal with a licensee later, if, you know, it's been 10 years. And now the product has, you know, six other features. And clearly a lot more value has been added to the code since it left the university. So I think the idea is you're looking at what additional value are they adding beyond what you've licensed to them? To consider dropping royalties over time? But it is a fair question. And it is one that that universities will discuss with licensees. There's a question on Speaker 5 37:31 average code changes 18% a year? How do you deal with that? Speaker 4 37:40 Right, so the question is, on average code, changes 18% Each year over time, how do we deal with that on derivatives? I will say that our license is not great in this respect, and we have not had problems with our licensees taking advantage of us. So you know, that the license is written more in our favor, you know, considering what's getting licensed today, because we don't know what's going to happen. We do have some language about, you know, we will renegotiate in good faith over milestones, which is not the same as you know, normal code derivation, but we would talk to our licensees, you know, in the future about changing the royalty, some universities will agree upfront to have some sort of, you know, de escalation of royalties. We don't do that, because we feel like you just can't predict I mean, when we never do it, I don't know, but we haven't. So, Speaker 2 38:39 Laura, as, for example, long term projects, you know, that ongoing for many years, and you've licensed it, and revenues flowing in, as graduate students come and go as postdocs come and go, the percentage that they've contributed to the code might change. How does University of Washington handle the revenue distribution for those licenses? Speaker 4 39:04 So most of the time, when we're licensing something, you know, it's fixed in time we're licensing it, we're handing it off. So we know who those cast of characters are. There are some cases where the software does continue to be worked on by by labs, not super common, but does happen. And any can get messy, right? It does mean that there needs to be some attention paid to Who are these people? And are they signing the right kind of agreement. So we make clear that we have title and we can move forward. And it can get a little messy because you know, sometimes we aren't always updated as much as we might like to be. But for the most, for most of what we do, these are fixed time points. So we have these check in points to be able to say to the research team, you know, do you need to re disclose this? You know, what has changed? Are there more people involved? Thank you. We're just I think in a finish on time. Speaker 3 40:08 So given the number of people that have basic copyright knowledge, I'm going to assume the same is true on the open source who's got some basic open source? Okay, good. So obviously another licensing models to distribute what you have under an open source under an open source. License. Is it coffee time? Yeah. Obviously, copy left, you'll hear that term thrown around a lot that is different than open source. And you'll hear some people use the two interchangeably. And that's just not correct. What I will say from, you know, primarily looking at this issue from an industry perspective. For the I started to say industry doesn't like copyleft. Industries, lawyers don't like copyleft creates problems. If you're depending, again, depending on what your goal is, with respect to why you're distributing something under an open source licensing model. Think about what that open source license is, do you just want to make it you just want to get it out there and make it easy for people to use? Do you want to control on You know how some of it is, some of it is used. And those kinds of questions will lead you towards, you know, a certain a certain open source, a certain open source license. Obviously, if one of your goals is monetization, you're probably not going to want to go with an open source model. Speaker 2 41:29 And that's right. So if you if there's around 100, open source licenses that have been approved by the Open Source Initiative, regardless of whether you want to do copyleft, or permissive license like an MIT or GPL, my recommendation would be to pick one of the six most well known ones, which are MIT GPL, 2.0, BSD, Apache, and Affero, GPL, and Mozilla. And then one thing to add is, you know, all open source licenses, the fundamental tenant is there are no restrictions. So once you license an open source software, you cannot restrict the users what they can do with it, even if you may be morally objecting to what they're doing with the software. So there is another set of licenses called open source like software licenses, then all universities have a flavor, because often we hear from faculty who say, we want this to be freely used and available for academics, but we want to limit it for commercial use, or for internal use. If you don't have one, I'm sure Heather and I would be happy to share a template. Another place where you can look up. Open source, like software is the Polyform project by Heather Meeker. They have come up with a range of open source licenses that are open source like licenses that have various restrictions, for example, restricted to small businesses, no commercial use for internal use. So these are some alternatives that you can think about around this issue. The other one Speaker 3 43:13 I'll mention is Creative Commons. I'll mention it only because I have I have seen it used in the software context, it doesn't. The Creative Commons licenses don't quite fit. But I've seen it done where I have seen some of the Creative Commons forms used more effectively is in that proprietary data set. context they have I think they have I think Creative Commons has one that is specific to data. And that one seems that one seems to work, work better. But the traditional Creative Commons are great for traditional copyright stuff. Software, it's just a bit of a round peg square hole. Speaker 2 43:48 I think it's creative commons zero, which is the one most appropriate for software, the other types of creative common licenses are more appropriate for creative works, internet. Yes. Could you could you use the microphone, please bear with me. I'm being recorded. Thank you. Speaker 6 44:06 So I see the question mark around monetization for open source software. Do you all have any experience or success tech transfer office was supporting open source startups at the university level? Speaker 2 44:18 So that's a nice segue to the next slide. And Speaker 4 44:24 I'll just say quickly, it depends on what you mean by that. So we had a faculty member moved to the University of Washington, he was mid career recruited in with some Jeff Bezos money to work on AI and he brought with him from CMU, some open source software and when he arrived, he started a company. So North neither the University of Washington nor CMU had any interest in it because it was open source software he had developed they're very successful was bought after a few years by Apple. I have a faculty member I will worked with, he wasn't sure he and his researchers weren't sure what direction they wanted to go, we contemplated going down a patent pathway for his software, he plays at the software hardware interface for basically customizing to machine learning. They ultimately decided not to pen and to build an open source community, and then go out and start a company. And they're newer, but they've gotten quite a lot of investment money so far. And you know, I'm optimistic they're going to be quite successful. But, you know, are they really coming through us? Or can we claim success? Man, you know, it's hard, right? We do claim some success, but not traditional. Speaker 2 45:47 So to add to that, I think data breaks from UC Berkeley was an open source success story, even though the tech trans office didn't get a disclosure for that, but it was completely built outside of the tech trans office. So that actually brings me to the next question, Laura, we have faculty who are releasing code open source, who could take the data, bricks approach, open source, release it under an Apache license or an MIT license, then spin out the company outside of the tech trans. How do you convince faculty to work with you? What are some strategies that have worked for your office? Speaker 4 46:25 So I think it's pretty easy to convince faculty to work with us. But that doesn't mean they're not gonna fall on open source pathway. Right. So I think they believe that there, there is value. So that second faculty member I mentioned who's got the machine learning kind of hardware software interface product. Two of his grad students participated in some of our entrepreneurial training, I think we actually funded them with some of our gap funding. So I think they believe that we add value and want to work with us, I think they were just looking for the best business model that they thought was going to work for the technology. And it seems to be working right, they developed an open source community by releasing the first code, open source, so we don't try to dissuade them from going down that path, we try to educate them about what their options are. And then you know, let them choose, I will make a quick side comment, we do not allow, to the extent we have control our researchers to release anything under a patchy or any other license type that includes a patent grant, because we are concerned it will unintentionally cross into somebody else's IP because those grants cross the university because they are general to the owner of the IP, not specific to the researchers. So be careful if you're going down those GPL, three, Apache type licenses. Unknown Speaker 47:51 Anything, anything else to add? No. Speaker 3 47:53 Okay, I don't think so. So what we've got up on the slide here is just you don't necessarily have to pick obviously, proprietary versus open source, you can have a dual licensing model, release for some release some parts or some, some part of the code for one purpose, and restrict that and then have another another commercial license for broader commercialization or other other for profit uses. So you don't have to, necessarily, Speaker 2 48:22 so just two comments before we move on from open to suffer. If in doubt, always pick a more restrictive license like a GPL. Because you can release a different version and remove permissive license. Once you release the code entered MIT license, it's hard to close, it's impossible to claw it back. Because if someone has downloaded the code, they have those rights under an MIT license. So if in doubt, advise your faculty to use GPL 2.2. And the other point, which I have now completely forgotten. Okay, if it comes back, I'll come back. Speaker 3 49:01 So we've we've figured out we can license this, we figured out whether we're going to release it under proprietary license or an open source license, how do you actually get it out there, depending on what the license or depending on what the software is, obviously, if it's a mobile app, you've got the mobile app platforms that are available to you. The Apple Store or Google Play, just straight website distribution. You know, I've seen folks that just have like a website, and you can literally click on the file and download the download the software, FTP feeds where you're pushing the code out platform polls where your licensee is logging into a platform and downloading it themselves. You can always go old school media. I dated myself already with the shrink wrap license, so we'll just stick with it. No old school CD ROMs that we all used to get thumb drives, you know, depending on on what the software is, and there could be hardware involved hardware or other other sorts of devices. The one that we don't have up here that occurred to me after the fact. Third party distributors Right, like if you don't want to distribute it at all, licensed to a license to a distributor, let them license it out, let them deal with the end users. You can still maintain ownership. But you get the benefit of having somebody else deal with the day to day hassles. hassles for you. Speaker 2 50:19 So, Laura, for the folks in the audience, you know, do you how do you distribute your software? And do you have any websites or technical providers that you use? Speaker 4 50:33 In our practice, express licenses means those high volume non exclusive licensing programs, not RT startup license, we've got a different name for that. So we do just changed about a year ago to a company called E loosen, and that's the platform we're using to push our Express licenses through. We also depend on our researchers to do the distribution. So they sometimes have websites where people you know, get a password, login and download. Sometimes if it's a small enough file, they're getting emailed to the licensee after we sign. In the old days, we used to do FTP, I don't think we're doing that so much anymore. Thank you. Speaker 3 51:18 So other alternatives, depending on again, depending on what your business goals are, depending on as you're doing your own diligence, but open source may be in the code as you received it, you may have some restrictions yourself SaaS, as a as an alternative to traditional software distribution, obviously, it's a very different business model, you're essentially providing a service at that point, software remains hosted behind your firewall on your systems, and your licensees. And users are just logging in and getting access to that software to use the software remotely, they're not actually getting a download a download of anything. And as I said, it's a very different business model, you've got hosting costs, you've got maintenance costs, you've got all kinds of bells and whistles that come with that. So may not be something that you want to consider. But it's always it's an option, depending on what the tool is and how it's anticipated to be used. Speaker 2 52:14 Laura, have you at University of Washington considered a SAS distribution? And do you have any examples or lessons learned? Speaker 4 52:21 Yes, we've got several different SAS kind of models. I would agree with Heather's comments that it's it's not a license, it's not a product. In that way. It's a service. And at the University of Washington, we have a goods and services policy that is different than the IP policy. And I would say there are some cases where it's problematic for us to get too involved in there are some cases where it's kind of a hybrid, where we've got some IPs, that that's associated with maybe a consulting kind of service. And it can be a little tricky to parse, what's happening in the department, you know, what's their responsibility, what versus what's our responsibility. And I'll also just generally say that researchers think they're ready to stand up a service, and they almost never are, right, that there is a lot involved in it. Researchers don't want to be available 24/7, you know, I would say, weighed very carefully and slowly into those conversations and, and movements. But we do have several different SAS projects. One large one is called the drug interaction database. It's a curated database of pharmaceutical data, there's a team of about six people who basically run sort of like a little business within the University of Washington. And, and I would say, it has been important and lucrative, but there are times including with this project, where the college it sits within isn't completely sure that it's a fit, you know, so it's definitely a different animal to be engaging in a in a real service project. One of my recent projects, wanted to do software as a service as the model because he didn't want to expose his source code to legitimate reason. I was worried about freedom to operate as she said, We'd we would have, you know, we're normally on the other side saying to our licensee, that's your problem, you figured out your freedom to operate. And I convinced him that instead, we would send out an executable desktop version, which means he's in the service business in terms of now somebody's paid us a lot of money to get a copy of his code. If it doesn't work, for some reason, he needs to be responsive, and we'll see how it goes. But I didn't want to go down the SAS model with him. So I dissuaded him in that case, but it's I think, a coming thing that I may get more questions about. Speaker 2 54:55 So while we own says, you know, I know several tech trends are pieces are thinking about machine and machine learning algorithms that they have trained on data sets. How do we distribute that without the data? So there, I've talked to a few trans officers who are thinking about sending it out. It's a zip file, it's, it's a train model, they send it as a zip file is a Python piece of Python code. They also send maybe 100 lines or data that they can use to test the Python code is installed correctly. How are they licensing it, they are licensing it either as a subscription. So periodically, they don't promise how frequent but they periodically as new data is input at the university site and the model is improved, they get an updated model. They also looking at perpetual license. That is the it's a license, you don't get any updates to the model, if they want an update and model the here there is an updated model. To think about these projects, I would say think about something that does create an easily distributable. So an example might be you have power lines and your power lines, you have data about whether the power lines in a zip code underground overground, how old is the power lines are the power, the tower, the power lines, or the wooden, the steel? So these kinds of information can be used to train a model to predict what's the likelihood of a power failure in this zip code. So that's a discrete project that can be distributed, for example, to insurance providers to think about projects that are discrete enough that you can distribute easily. And then you can push out periodic updates to the model. And Todd, so you had a question? Yeah. Speaker 5 56:46 Todd here, Emory University. And Laura. I was in Oregon and Washington in the early 90s, and 2000s. So I know your Maverick group? Well. My question is this, and I was hoping you guys would answer it, because we're at the front end of this is sort of research contracts that we're entering into. And it's not just research contracts. Now, it's contracts, at least at Emory healthcare for with groups like Accenture to help us create patient centered dashboards, with Betty Ford Foundation to bring some of their clinical curriculum into our hospital. Some of the most robust IP sections I've ever seen are falling into these agreements now. And a lot of it hinges on copyright rights. And at the center of much of that is who gets to own the derivative rights. So what I want to know is what you're doing, it's your institution to and and who owns derivative rights, I understand the copyright law, as long as we have permission to create derivatives allows us to own them unless we contract them away. For we're having a really hard time retaining ownership of derivative rights. What should we be doing in those contracts? You know, is it okay to be asking to retain ownership of the derivative rights? And how do you? When do you and don't you retain ownership of those? Speaker 4 58:04 So at the University of Washington, are offered office of sponsored programs has the ability to go to the researchers in it within certain restrictions that I think sometimes they exceed. If they get faculty and departments sign off, they can, you know, go beyond what is sort of the standard offer. I will say my experience is that at least I think our office is just my opinion, is underfunded, not not my office, the office of sponsored programs. And often these conditions are showing up in industry sponsored agreements, we get a lot of federal funding in the University of Washington, that's the really big dollars. So I think there's a lot of pressure on them to just get the deal done. And faculty sometimes just want the money, right. And so we see all kinds of conditions coming through to us. So if we have a relationship with a researcher, they might come to us ahead of time and ask us for our opinion on some of that language. Occasionally, the office of sponsored programs will try to sort of make it our job to negotiate that piece. But often, something just gets signed, we see it downstream and it is what it is. What's your advice? What's my advice? You were doing it? What's your advice? Try to have enough of a footprint on campus that researchers will come to you or will understand what they are giving up in those agreements. And then if they give them up, they give them up break. You know, they're kind of driving the force of the university not not tech transfer. Speaker 7 59:49 The angel from Fred Hutchinson Cancer Research Center. So we talked about the distribution model. How do we distribute software without giving up the ER or without exposing the source code and also allow us to continue to improve. So one common way to do it in the industry was to build a field way, a connection, like build a connector that allows the partner to call the result while not accessing to the source code. So API was trying. So I wonder, well, perhaps hasn't run into a situation like that. But I'm wondering if any one of you have run into a situation where API would be used and how that like allowing access to the API can't be reflecting. Speaker 4 1:00:41 So then you're back in the service business, right, so you've got this proprietary code behind your firewall. So you have to know that research team has to be willing to take on that responsibility. Not many research teams are interested or capable or even stable over time, maybe the hunch is different, because you've got more stable staffing than a lab with transient grad students. You know, I wouldn't say never, but I'd say you're you're waiting into that service business. Are they really prepared? And can they pull it off? Speaker 8 1:01:17 Thank you. Yeah, so I just wanted to go back to the comment early comment about who constitutes an author. You know, when I think about software, there's someone who has the idea. And there's someone who maybe draws a block diagram, then there's someone who writes pseudocode, then there's someone who writes like a proof of principle. And in one case, we had the researchers did the proof of principle worked well, they wanted to start a company. And they said, you know, this is a proof of principle, this is not production, quality software, we're gonna have to rewrite anything, we don't really need you, we're gonna go off on our own. So you know, as you think about authorship and derivative rights, you know, how do you navigate these sorts of issues? Where did WHERE DO WHAT IS WHERE DOES author start? And where do derivative? Derivatives begin? I guess. Unknown Speaker 1:02:06 Is that me? Speaker 3 1:02:09 It's yeah, it's, it's, it's a great question. From the perspective of, of actual copyrightability, and copyright law, just the ideas, the suggestions are not protectable. It's not until you get that it's not until you get that code that's actually written, saved, stored. And so it's that person who is the author under under the copyright, under the copyright law now, how that may fit into your institutional policies about what they may say about, you know, in any sort of contributions is perhaps a different question. But from a straight copyright perspective, it's it's literally the author is the person who's writing the is writing the work, think about it. Some of these conversations are easier outside of the software context. So you know, pick famous novel, X, you know, famous novel X, that author may have had somebody in the background saying, hey, you know, it'd be really cool if you wrote a book about blah, blah, blah. But until you actually write it down, it's not, it's not. And that's a difficult conversation with somebody who's gotten the funding, who's, you know, who's been deeply involved in the project, but didn't actually write anything. And Speaker 4 1:03:23 I would say your question also gets the value, which Heather's got a slide coming up in a couple that talks about valuation, right. Maybe your code isn't that valuable in that circumstance. And Speaker 2 1:03:34 in that example, the flowchart creator will have copyright to the flowchart. Correct. But the code writer would have the copyright to the coder code, right? Unknown Speaker 1:03:43 Yeah. Because you would, yeah, it's a great point. Speaker 8 1:03:46 What about like, the derivative works like we don't need you are going to rewrite everything? Speaker 3 1:03:53 Yeah, so the specific question, being with that, if if you don't need the original flowchart, creator anymore, you can just rewrite everything from scratch, meaning that as the if the institution is the owner of the copyrights, then you've got that you've got that right to create those derivative works. And yeah, that's something that you can absolutely do. Speaker 9 1:04:17 Got a question for you. A lot of researchers that we deal with are federal employees. And so that's a problem with copyrights that federal employees can't get a copyright. So what do you do for protection in a case like that, and then also, going back to your question on patents with the open source? So if the machine learning model was created, at least in part with Apache, are you recommending that you not try to patent it or Speaker 4 1:04:49 not? So Apache is fine if we're receiving, right? We're happy to have somebody else give us their path. Right? Right. We're Apache is problem in codec is when one of our researchers wants to attach it to his or her code. Because it contains it says something like, to the extent you have a patent that is needed to practice this software, you're granting us those rights. And our concern is not so much for that researcher, because generally we can do a better job of assessing whether that researchers other patents, should they have any will be affected. But what about somebody else we're not thinking about because we all have these broad diverse patent portfolios. So we just don't want the specter of somehow inadvertently granting rights to some other person's patent by allowing software licenses that contain those patent grants. Speaker 2 1:05:44 So I have a slightly divergent view on that. So when I get asked about open source, and patenting, I always say, if your intention is to open source it, we are not going to patent it. And that's because even if we were open source licensing, an MIT license, yes, it's silent on patent. But I am concerned that we could enforce the patent because we openly given the software way. You don't get two bites on that apple. So I, you know, we've given the software, they've used it, and now we are going after them for patent infringement. So I, you know, I have a conversation early on with my faculty, if the Open Source Path is what they want to do, then we shy away from patenting. There Speaker 4 1:06:28 one other comment on the federal employee thing I did have a technology that had was, I don't know if we technically call it a joint work, maybe Heather could comment. There, there was a federal employee, but there were also University of Washington employees. So they didn't have copyright. But we had copyright. And we tried to figure it out legally. And it seems like a murky area. But we decided in the end, we had copyright in it, the federal government didn't and we could enforce our own copyright. Speaker 3 1:06:58 I will reiterate, it's a murky area. Can we leave it at that? So the next slide, we've got up here, some of those diligence questions. So you get a disclosure, this stuff comes across your desk, we've we've you know, we've beat the horse about authorship versus ownership and how you and how you get there, I'm not too sure there's much more we need to say there. Just to just to add to our bullet points here, because it didn't, we didn't get it on the slide, that data privacy issue, right. And in terms of the ownership of maybe some of the back end data that's used to drive and develop some of these some of these things, something, something to add to that list of list of issues. Those open source restrictions. And so just as we talked earlier about how we might be distributing something under an open source model, looking at the software that comes across our desks, and looking at what open source is in it, because again, that may also restrict us from what we can do to commercialize anything that that that we see, you know, Word to the wise, build your policies to have people track the open source as it's going in, not having to go back after the fact and figure out and figure out what's what's in there. And you guys probably have some more real world examples of when that's gone badly. Speaker 2 1:08:18 But so, on that, Laura, do you at University of Washington use software to figure out what open source software has been used in project like Black Duck, for example? No, Speaker 4 1:08:30 we have a spreadsheet that we send to researchers, and we ask them to provide us with a link to the license or in some cases, the actual license if they have the text. And then we do the rest of the due diligence. You know, we interpret what that license means in the context of how we're trying to commercialize. And then we put all the risk on the licensee, we say, you know, you get it as is that, you know, we're disclosing everything that we're aware of, you know, it's on you. Speaker 2 1:08:57 So what happens if the spreadsheet comes back with a problematic license? Speaker 4 1:09:03 They, the researchers can choose to rip that part out. And we can license without that part until the licensee you have to find a piece to plug in here. They can rewrite the code, you know, as long as they're not, you know, infringing on that third party code they took. Yeah, so that's what we do. Talk. Speaker 5 1:09:24 Yeah, I'm sorry, can I go back and just ask you to beat that horse one more time on authorship and ownership? Because I thought he asked a question and you answer the simplistic situation. And maybe that's because the answer even for the more complicated situation, but if I'm a physician, and I've got somebody who's writing software code, and I want to help predict sepsis, and that might occur, and so I'm going to collect data in different ways. I want to write some code and I want to analyze that and I'm going to ask you to do that for me. You have no idea about what sepsis even is. And so now you come back with some code next See? Yeah, no, it needs to do this a little bit differently it needs to take into account this situation. Am I still not a copyright author if I participated, even to that degree? Speaker 3 1:10:10 I think I'm those facts, probably not. Because again, it's the actual expression of the idea, not the idea itself. Speaker 5 1:10:17 Yeah. And so there's never a case where unless you're writing the code, no matter how tricky it gets, in figuring out how to make that code, do what you need it to do, that you become a copyright author. And remember, Unknown Speaker 1:10:27 I said Never say never. But 99.9%. Speaker 4 1:10:33 Okay. And I think that, that that's a common issue. That right, a faculty advisor does not like the idea that they are not an author on code that's coming out of research and valuable contribution. Right. Right. And so, you know, there are ways that, at least in our system, we can include other people that are not technically authors. And, you know, sometimes there's this sort of fuzzy vision of we may, you know, we may record them as authors, because we're trying to set up a circumstance where everybody's winning. But there are also cases where we can allow them to add somebody else who has been really important to the development of that we had one project that had a great windfall, and kind of an administrator had been really pivotal, because the faculty member needed an administrator to make sure he got his stuff done. And she ended up I think, getting like a 40 or $50,000 payout. So, you know, I think there can be ways around that. And it is a real issue, I think, for most of us. Speaker 2 1:11:40 So I think, to add to that, we addressed that by allowing people to be recognized in the revenue share, rather than as authors. Speaker 8 1:11:53 So my question about the derivative was, the company wanted to evade their obligation to the university by rewriting the code. And so my question is, when does a derivative start? If they if they take the ideas and completely rewrite the code without copying anything in the original software, is that separate? That's Speaker 3 1:12:13 when you start to get into a copyright infringement analysis? And you don't have to? It's complicated. No, you you can't necessarily get around it that way. If you when you look at something from a copyright infringement analysis, if you have substantial similarity, you can still have copyright infringement even though so that, in that particular instance, with those with those facts, I think you would probably still have some claims, you can't just completely look at a work and say, Oh, this is great, I'm just going to completely rewrite it. Because your intent is to do just that, to get around the copyrights. And that's not okay. So we've only got a few minutes left. Speaker 10 1:12:55 I had one one follow up question. Sure. The university often have huge teams of students over long period of time that are contributing to this, so you could end up with hundreds of craters. How do universities and other institutions manage that? Speaker 4 1:13:15 If you're talking about hundreds of creators, well, one, I've never seen hundreds of creators except with one giant project that we manage that's multi multi multi institution. But, you know, if you have a lot of students, first of all, I'd be asking, are those students employees? Are those students undergrads talking about undergrads, we typically don't own their work. And that would not be something we would probably spend much time on. If it's a, you know, series of grad students over time, and, and we caught it probably early enough, you know, we would try to help manage through those changes of, you know, cast of characters. And it can be tricky to kind of to the earlier Drew's question, how do you share the money over time, and that can, you know, be a little more complicated in a situation like that, but, but if we're truly not talking about hundreds, we would, you know, we'd work with them, like we'd work with anybody who would just be more painful. Speaker 2 1:14:14 In one instant, we have actually gone back to the authors and renegotiated the percentage that they receive. So over time, the old authors received less, and they agreed, we had to agree get them to agree to it, but we didn't renegotiate their distribution of royalty. Speaker 3 1:14:32 So I think we've got like three minutes to cover valuation. So we've put some ideas up here. At the end of the day, is there any way to value these things from a license perspective? Not really, I mean, any transaction requires a willing buyer and a willing seller, whatever price you agree at is what your market price is. The question is, how do you start those conversations? These are some of the data points that I suggest people look at. What did it cost you to develop like, did you have to buy equipment did Do you have to like, what what investment did you put in? How long did it take? How many, you know, how many worker hours were there? cost of replacement cost of replication, similar, there are comparables, databases. I know there's a vendor downstairs, that purports to be able to do some of that, and give some ideas for licence fees. And then the other one that we came up with talking earlier, cost of maintenance, if it is something that you're gonna have to maintain on an ongoing basis, what are those future costs? going to look like? So I, Laura, I don't know if you have any other thoughts or suggestions on ways to get to that valuation? Yeah, I Speaker 4 1:15:37 think Heather summed it up nicely. That, you know, there's no magic formula here. And sometimes I felt like I'm just pulling a number out of the air. And sometimes I am pulling a number out of the air. We have enough deals at the University of Washington that at least they're anchored in some past history. But as Heather says, you know, there has to be a willing buyer. So you know, if you've got somebody who's saying, I don't think your stuffs worth anything, you know, I'm gonna go do it myself. If TRULY, they can read our published paper and go do it themselves, good on them, go for it. But if you need our code as a reference code, then you have to come to me and we have to agree on some value for it. And you know, that can be a lot or a little. Speaker 1 1:16:31 Great, thank you, George, Laura, and Heather, for such a great discussion. I know there was never enough time at the annual meeting. And I'm sure we'll run into the same thing here today, attendees. As a reminder, please feel free to submit your questions through the q&a. Or you can raise your hand if you'd like to ask your question aloud. And with that, George, I will turn it over to you. Speaker 2 1:16:52 Thank you, Sammy, and thanks to all the folks who were able to join us this morning or afternoon, depending on where you are. And thanks, again to Laura, and Heather, for joining us for The Live q&a. So maybe I'll start with the second question that came in, which is to Heather, which is what would you recommend regarding registering the copyright for software codes, especially when it is likely that code will evolve pretty frequently? Speaker 11 1:17:23 Yeah, absolutely. The So recall that a copyright copyright exists as of as of a moment in time, right. So if you are going to be if your code is going to be evolving quickly. Then once you register that code, you can certainly update those registrations later. But it The point being that as as your code evolves so quickly, then you get further away from what's actually registered, and you lose some of that benefit of that registration. Because what's captured in that first registration, as of that moment in time, is what's covered by covered by the registration. So generally, what generally what I have folks do for for sort of key software assets that do evolve, that do evolve fairly quickly, they will just sort of establish a standard of updating the registration quarterly or updating the registration once a month or updating the registration with with some periodic frequency, the other way to go about it is to look at to sort of look at your updates, and wait until you have a significant update. And then and then update your registration from there. But the to the point, I think, at the heart of the question, because the purpose of the registration is to have the ability to enforce that copyright as it exists. The as you when you develop something and continue to develop something so quickly, you do, I think lose the benefit of that registration process. And you just have to do it more more frequently, which, you know, obviously creates additional process issues and additional costs. So, but best practice, if you have that kind of scenario, and it does make sense to register your copyright is to put one of those processes in place to update it periodically, whatever that frequency may be. Thank you. Speaker 2 1:19:23 Laura, I'll start with you with this question, which is, do you have any recommendations for SAS license templates, guides or resources that you've used? Speaker 4 1:19:37 Anybody can go to the University of Washington commotions website and look for our Express licenses and on there. You can actually look at the licenses for the drug interaction solutions database, which I mentioned in the talk. And so you can see what one of our our service agreements actually looks like. It's a subscript chin agreement. I was, I don't know, a while ago contemplating some different kinds of service terms. And I watched a video of Heather Meeker that I just found on YouTube talking about a software service agreement. And at the end of it said, If you emailed her, she'd send you her template. So you could do that, too. We have we don't do a lot of service kind of agreements, because that's actually, as I mentioned, the goods and services policy of the university rather than tech transfer, but you know, we've, you know, dabbled in it. Unknown Speaker 1:20:39 However, do you have any ideas? Yeah. Speaker 11 1:20:40 The other thing I'll add to that is if is if your legal office has access to Westlaw. Westlaw has an PLI as well, practicing law Institute, they have forms and templates in their databases. They're pretty generic forms and templates, but those, but those those are out there. And if you do happen to have have access to it, counsel, information technology, counsel, information transactions, counsel, they most good it lawyers are going to have a good starting point. Speaker 2 1:21:23 While I was at university, sometimes I've also found templates in law Insider. And while I was at Purdue University of Chicago, and at Northwestern, we had subscriptions. So your institution might have subscriptions to some of these services. And we might not know it at a tech transfer office. So I would check if you have subscriptions to some of these services. And the next question arise. How do you balance the desire for computer scientists and researchers who want to release their code open source, but also reserved commercial rights? Yeah, Speaker 4 1:21:59 that's a great question. In addition, and maybe this is what the questioner was thinking of, too, we're starting to see more journals requiring some sort of, you know, provision of the code to demonstrate that, you know, their research is, you know, worked. So, always, like, want to have a conversation with that faculty member early on. So for me trying to get in front of that, and make sure the faculty know who I am, is a big struggle, particularly with computer science, because we're hiring so rapidly, I feel like, you know, it's hard for me to even keep track of who the faculty are. But it, you know, in the spectrum of, you know, if I was in charge, and we had tech transfers, our primary goal, which is almost never the case, you know, we would want to keep things more buttoned down. But then, you know, in the spectrum going from, you know, they don't know, all the way from where I am to where they don't really care that much. There's lots of space. And we can sometimes release under an academic use license or write some sort of custom license for a journal, if it's just, you know, providing code for the journal, and keep our other rights. You know, anytime you expose the source code, there's some risk there. So I have some researchers who put their code on GitHub, we put an academic license up there, but of course, the source is exposed, and there's, you know, some risks that could diminish the value in a commercial license later. It's not always the case. So I just want to have a discussion with those researchers early on. So we make sure we are we are picking the right license that really supports the goals they want, sometimes even sophisticated researchers think they understand what these open source licenses are. And surprisingly, they do not. Thank you. So I Speaker 2 1:24:01 know we touched on this during the session, Heather, but should you always wait to register copyright? Only when you plan to file a lawsuit? So why why bother with a copyright for software? Speaker 11 1:24:15 Yeah, and it's it's a it's a valid question. Because again, when you do when you do register the copyright in your software, you have to make at least part of the code out there, out there and available. Really the only the only value you get, I shouldn't say the only value you get. But yes, it when when you're looking at whether or not to register a copyright in your software. The key driver is whether or not you anticipate needing to sue on it. You can wait to file that registration. At the point where you realize you do need to file a lawsuit. You lose some of the benefits of the registration in that instance and what I mean by that As if you've if you've filed the copyright in your software before and infringement starts, then that will then doing so either before the infringement starts within or within three months of publication, you get additional, you get additional remedies available to under the Copyright Act, you can wait to file that copyright registration, when you see that, when you see that a lawsuit is going to be necessary. And you will still be able to get into court, but you're not going to get those additional remedies available to you treble damages and things, attorneys fees and things, things of that nature. So you certainly could make the business decision to wait particularly. And as I see some of the other questions popping up here. Particularly if if you are going with a SaaS model, for example, such that your code is never distributed. In that kind of instance, the likelihood that somebody is going to be able to get to your code to reverse engineer it to copy it to modify it is less. And so in that instance, I think the likelihood that you would, that you would perhaps need just to on it is lower, and that might drive some of the decision factors on whether or not you want to register for it. So things it's a fax of that nature that would that would go into that analysis about whether or not it makes sense to to register a copyright in any in any software. Thank you, Speaker 2 1:26:29 Laura, what about inventions that are used open source codes, in creating their proprietary software? What are some key considerations when you commercialize those types of innovations? Speaker 4 1:26:45 I'm not sure I understand the question to the extent I get it, I think they might answer is, it's fine. If somebody uses open source code, and they create an invention, that's not the open source code, right. It's something else that uses it. So I wouldn't be concerned at all about my patent rights. If there was some new invention there. You know, I might be looking at what are the what's the open source license mean, if there's software that goes along with the invention that I would like to commercialize? Speaker 2 1:27:19 And I think you made this point where you have an Excel sheet of, you know, a lab and, you know, concerning open source licenses. So I think it goes to that point to Laura that, did the inventors or creators use some copyleft license, then which would mean that if you were going to license it, you would have to read those under a copyleft licenses were Speaker 4 1:27:44 right, yes, absolutely. I would be looking to see if there was some sort of condition than the open source license that reached through to my potential invention. If, you know if, if that was relevant in that case. Speaker 2 1:27:59 Thank you. Um, Heather, this is a question from Bill Jones, and derivatives from initial court implicitly covered in some instances, given the consideration for substantial similarities in copyright infringement. Is this ever addressed in a license? It seems it may be difficult to identify when, when a true derivative, which is free from original work of authorship has been created? Speaker 11 1:28:29 Yeah, so I'm, I'm going to take a stab at answering this one. And we'll we'll see if this answers your question or not, but recall that when you when, when you look at what a copyright is, you've actually got a bundle of rights, right, the copyright owner, the copyright owner actually has the exclusive right to make derivatives no one else can make derivatives without getting a license from, from the original copyright owner. And so if in fact, the copyright owner grants a license to you as as the licensee to make those modifications to make derivative works, then you're going to have that license and you're not going to be in a copyright infringement. Kind of kind of situation because your your license is in fact going to be a defense. A defense to to infringement. If you do not have a license from the original copyright owner to create droop, that derivative work, then technically you're infringing the copyright and any derivatives are sort of forfeited as a as a matter of law in because you violated the the copyright owners exclusive copyright to make those derivatives. So I don't know if that answered your question, but if not, we'll then feel free to pipe back in. Speaker 2 1:29:49 Thank you. Um, so there was a question on open source licenses that are better or more standard. And so I can take that And I think there's one open source license you take really depends on your business use. So there is I don't want to say there's better ones, there's less better ones. But of the 100 or so that the Open Source Initiative as approved, six probably account for 99% of all open source licenses out there. So I've had this experience where when I was writing code, I would look at a piece of code and I look at the license. If it was MIT license or a BSD license, I know the terms. If it was one of the less used ones, there is, I would say there is an activation barrier for me to now go and look at it. So think about when you pick a license, do you want people to quickly adopt it, if there is something that they have to do extra, it's less likely they're going to look into it. Another thing that I've noticed is, sometimes people put codes and they have using an MIT license, but they don't actually mention that in the text file. That's a good practice to mention it because it avoids the your developers or whom you want to encourage use the code to have to figure out what codec so there is this kind of activation barrier to use the code. So if you're looking to wide adoption, I would pick the ones that are most widely used, which is MIT BSD, GPL, Apache Mattila. And then make sure that you pick those because they are well known that terms are known. And then that leads to wider adoption. Speaker 11 1:31:40 The other thing that I'll add to that, coming more from the from the software industry side of the spectrum, lawyers hate GPL. In Mozilla, as well, like Eclipse is another one. There's there's the the one the ones that George mentioned, MIT BSD Apache, like those, those are the three that when I see those as it counsel, I'm like, Okay, I know what they say they're clear. They're pretty low risk for us to use in a commercial setting. When you start for those of you who have not read the GPL. It is, it is an interesting read. It is far from clear, far from concise, and interpretation of that in any given situation is not straightforward. And so Georgia, your comment about a cost sort of barrier to entry, it's a great, it's a great way to look at it when I when I see something from a commercial perspective that has GPL it's a great, I have to think about it. Now. If you go with something that's that's MIT, BSD Apache, obviously, those are very permissive licenses, which may or may not match the goals that you have for the for the software. But it is a lower, lower barrier to entry. Thank you. Speaker 2 1:32:55 I think we have one final question, which is actually on patents. But how valuable is a patent covering software at this point? Do companies licensed software care if their patents on those software? Maybe, Laura, from your licensing experience? Speaker 4 1:33:15 Yeah, I think the answer is it depends. If the project is going to go into a startup, I think often patents do help startups, sometimes I feel like investors don't even seem to care what the patent application says. But having something seems to give some credibility to the team. So you know, and some patents covering things implemented in software can be very foundational and important, too. And then there are cases where, you know, you can have a weak patent. And, you know, in my world, I think about, if we get revenue from this, will we ever, you know, even recover our patent costs, much less push any money back to the researchers, right. So that's more kind of strategic decision about, you know, what value that might have. So I'd say it depends. Speaker 2 1:34:09 Thank you. Are there any folks who want to ask questions live Speaker 1 1:34:18 I'm not seeing any hand raises. We can give it we can give it a second is there any while we wait for any closing hand raises? Any kind of last words of wisdom the three of you would like to share before we close off of keto? I know we've covered a lot but kind of quick one liners of what not to forget. Speaker 2 1:34:38 Um, Laura and Heather. And any anything new since we, you know, we met in New Orleans, any new trends that you're seeing both from an academic setting and from your clients? Unknown Speaker 1:34:54 Not Oh, no, it's all you Okay, Speaker 4 1:35:01 I feel like we just continue to see, you know, AI in everything, it feels a little bit like 20 years ago, where.com was on everything. Now AI is in everything. So figuring that out data, those seem to be, you know, the new, new things that are sitting on my desk. And Speaker 11 1:35:22 the other thing that I will add to that is, you know, everybody, everybody wants to do a SAS these days, I think the last it's been a long time since I've done sort of a traditional on premise software license. And the day reminding folks that it is, in fact, a service, and, and it comes with ongoing obligations, and how are you going to manage those? And how are you going to manage those costs, it sounds cool and sexy to say you've got a SAS, but it comes with a lot of baggage. Speaker 2 1:36:00 And so there was a question on would large companies find patents valuable? So I can't comment on invidious position on that, even though I worked there. And because I work there. But in general, from a university perspective, we've certainly found patents to be valuable, right. And we know that certain universities have sued large companies for patent infringement. So again, what's the value of the patent? It ultimately is a bit, you know, there's a cost to patenting it. And so what's the value you gain? And how is it foundational work that you think is going to have many applications? So think about those as you think about patents, and whether companies would value those patents? Speaker 1 1:36:49 So I think with that, it looks like we have all of our questions that have been submitted. On behalf of the whole autumn team. I'd like to thank you, George, Laura, and Heather for joining us today for such an informative discussion. And attendees. Thank you so much for joining participating, submitting your questions and helping us have a lively q&a today. Transcribed by https://otter.ai