IMS Elite Expert Sam Rogers discusses the impact of global content creation on infringement disputes and the importance of YouTube’s Content ID tool.Listen above or read the interview transcript below. (Part 2 of 3)

Hello and welcome to the IMS Insights Podcast. I’m your host, Adam Bloomberg.

Today, we’re speaking with IMS Elite Expert Sam Rogers about the importance of proper investigation, common legal disputes, and the YouTube Content ID tool.

Sam Rogers is a social media industry and instructional technology expert. He provides strategic learning, marketing, and content creation services to global companies such as Google, Deloitte, ADP, Capital One, and Robert Half International. As a recognized expert in the field of Learning & Development, he routinely works with subject matter experts in technical fields to distill their expertise into training that meets organizational needs.

Adam Bloomberg:

So, let’s talk a little bit more about your work. During discovery, what are the issues or problems you’d typically run into?

Sam Rogers:

Well, in discovery, typically, there’s a huge volume of content that people need to make sense out of, and they haven’t yet. And by this, I mean there’s a CSV file full of a bunch of numbers and a bunch of references. And even people who really know how to do things with pivot tables and Excel spreadsheets and all of that, they don’t necessarily have the underlying data model in their head to know what’s important, and what isn’t, and how these things relate. So, one of the first problems that I’m generally tackling is just trying to figure out for this, say set of videos or something like that, how were they used?

Sam Rogers:

Or, for these users who are in these places, how might the YouTube algorithm surface content to them? For this body of content here, that has these plays and these places, how would that be benchmarked against other content, other channels, or other creators? So, some of that is bringing information that’s just my knowledge about the temperature gauge of what else is going on in the world. But, a lot of it is really diving into the details, doing some data analysis to be able to begin to look at what the valuations of these numbers might be.

Adam Bloomberg:

So, what are some of the common legal disputes that develop in your industry? And then, what are the types of cases that you’re seeing resulting from that?

Sam Rogers:

Well, as I’ve mentioned, the arguments that surface to litigation are generally the monetary ones. So, in terms of what people are disputing, there’s basically three different kinds of things. So, one is that YouTube has been used to distribute content, and it probably shouldn’t have been, or it shouldn’t have made money off of this or distributed money to the people who uploaded that content or something like that. Then, there’s also the angle of the search algorithm; how it is that YouTube surfaces content? How likely is it that this person in this place saw this video from this person over here? That can be relevant. So, in one case, although I have some statistical ability, I’m not a statistician, so I don’t generally publish the numbers, but I run the numbers. And from that, I can help the counsel tell a more cohesive story with things that people can relate to, about what these use cases really look like, how likely it is that this might happen, or that has happened.

Sam Rogers:

And to set a value on the plays of YouTube videos or the instances of something that’s been uploaded to this global platform. Remember, we’re talking about things in the context of US law here, as in expert witness in the US legal system. But as a global platform, YouTube actually has to account for every possible law that could exist around the world, in 190 countries out of 195. They have to have some mechanism where people can control and distribute content via that platform. So, it’s quite a complex thing on the back end, especially in those cases that are global in nature. But certainly, there’s a lot of standard US terms that come up in terms of what is copyright. What is fair use? What is transformative? What is the content ID system? How does the DMCA Takedown process work? All those kinds of things.

Adam Bloomberg:

It’s interesting to hear you talk about what goes on behind the scenes. I have a subscription to YouTube TV, I just got it a few months ago. I love it. And it’s so interesting, because when you open up the app, you choose the user. And in my household, we have four users, and you go in and then, you basically go left or right. On the one hand, you can look at all the content out there that’s just free to the world on YouTube, or you can go into the TV section. And every time I turn that on, I’m thinking, “Wow, there’s so much out here that I haven’t even gotten to the TV part.” And there’s so much here just on YouTube, but I would assume, like most of the average consumers, I’m not thinking about how did this content get here? What is this content? Does the person who owns this content put this up, or was it somebody else? So, let’s talk a little about that. You’ve mentioned YouTube Content ID. So, why don’t you tell us what that is? And maybe, we’ll start there.

Sam Rogers:

Sure. It’s a system that’s governance goes back to the Digital Millennium Copyright Act, which was the most recent substantial revision to copyright, which was the 1998 Act of Congress. YouTube comes along in 2005; it’s what gave rise to their ability to create a global platform. It was being able to define what the responsibilities of a provider are, like that, and what Content ID is, which did not exist in the early days but has come to much more prominence in the last decade or so. It’s a way of identifying content that’s getting uploaded to YouTube and matching that content to other copies of it that may exist already or that may exist in the future. So, there’s over 500 hours of content that are uploaded to YouTube every single minute. So, in the time that we’ve just been talking, how many hours is that?

Sam Rogers:

It’s no wonder that you’re looking left and right. You’ll never get over to that other side because there’s just so much out there. And a lot of that is infringement that YouTube is preventing from being distributed to the world. And they have an obligation to do that. You and I uploading cat videos, we might not understand all the ins and outs of the legal process or what constitutes copyright. Maybe you and I do, but I don’t know your cousin doesn’t. So, I might attempt to upload something that I don’t actually own. Like here I am at the concert, I’m just pointing it at the stage, this is mine. Well, is it? That whole performance that you’re witnessing, the songs that are playing, those aren’t actually yours. So, YouTube’s content ID system is able to recognize things that are recordings, be it audio or video, and actually match them to other audio or video content.

Sam Rogers:

Even if it’s been cropped or the color is changed, or it’s flipped or backwards or even if it’s in another key or pitch shifted or all of that. That system has a way of spotting what’s happening and matching it to this existing corpus of content, such that they can provide a way for content owners to express their wishes through their platform. So I, as someone who has a YouTube channel, I can upload a song there that I’ve recorded, and I can say, “Oh, well, I want this to play here, but don’t let it play over there.”

Sam Rogers:

“I can upload this video here, this is my music video, but I’m going to take it down now because I’ve decided that I don’t want to have that music video up anymore.” I have the ability to do that, and when I do, it’s no longer displayed in YouTube. So, I have these tools through the YouTube interface. Content ID is a lot of those same kinds of things, but it’s a much more amplified backend tool set that you and I uploading the cat videos would never see. So, say if I’m National Geographic or something, and I have a video clip of a cheetah running, and you want to use that video clip of a cheetah running in your video. It’s actually their content, and content ID is what matches that clip of the cheetah running and allows them to apply a policy to it without even having to see it.

Sam Rogers:

Someone at National Geographic doesn’t have to watch your video and figure out what kind of thing it is. They have the opportunity to do that, but with so much content getting uploaded so much throughout the day, it’s just not humanly feasible to do that. So, being able to apply proactive policies that say, “Okay, you can have five seconds of the cheetah running, that’s fine. If you put five minutes of the cheetah running, I get a cut. I’m going to monetize this video, and there’ll be some money that happens. If you’re going to upload an hour of the cheetah running, well, now we have a problem.”

Sam Rogers:

Now, I can’t have all of this display because there’s other, maybe rights or agreements that I have in this jurisdiction or that jurisdiction. Well, now it’s a different game. So I, as a Content ID user, can apply these policies proactively to content that I have uploaded and that I can manage. And then, YouTube, as the distribution mechanism, can apply those policies in the appropriate way so that they play to the people who are intended. And they don’t play to the people who are not intended, not by YouTube, but by the actual copyright holder.

Adam Bloomberg:

So, what are some of the criticisms of that technology?

Sam Rogers:

Well, there are plenty. As with anything, whenever you start applying machine logic and things like that, there are false positives and false negatives. It’s analogous to your email inbox, right? There’re some things that go to spam that really, most of it needs to go to spam. You don’t need the bank offer from the Nigerian prince again. But there’s some stuff that sometimes winds up in spam, and it shouldn’t go to spam. Where’d it go? That wasn’t spam. And then, there’re some things that do make it in, right? They make it into your inbox, and you’re like, “Do I know this person? Wait, I don’t think I do. Should I click on this link? It’s spam.” So, those rules that I was describing for content ID, it works in much the same way, where you can say this amount of time or for this kind of content apply this policy. But there’s always going to be things where it’s not applied evenly, or I should say, where the rule is applied evenly, but the content, it doesn’t quite fit.

Sam Rogers:

There are also very commonly multi-claim situations, where several different entities may have claims on one piece of content that has been uploaded, multi-party disputes about that. Reasonable people may disagree when we get into something as simple as a song, there are different kinds of fractional ownerships that exist, different kinds of rights that exist, different kinds of agreements, which may apply in different jurisdictions. It’s not always clear and spelled out exactly what YouTube should be doing. So, it’s confusing at best to be able to straighten it all out. And in terms of the criticisms that they get, largely, the most vocal criticisms come from the creator community. And I can share a story myself, if that’s all right? So, as a musician, back when music happened on shiny plastic discs, I made a few of those, and I signed some agreements about what can happen with that music.

Sam Rogers:

And it was uploaded to YouTube, not by me, but by someone I had signed an agreement with. Well, to YouTube, they know that they signed that agreement with that entity, but they don’t have any way of linking that with me. So, when I perform my song from my album, and YouTube flags it as being infringing, I’m not infringing on my own thing, but YouTube can’t tell the difference. So, creative people who don’t have an understanding of these subtle distinctions get really mad, and they start doing things like not just complaining, but pressing all the buttons they can throughout YouTube.

Sam Rogers:

So, sometimes they accidentally initiate legal processes such as a DMCA Takedown, that really are an inappropriate use of that function. And then, people get all confused, “Oh, I thought this was fair use.” Fair use? Good. Fair use in the US, where fair use is a legal concept that applies, but not fair use over here in Germany, where that’s not a thing. So, your video can do that here, but it can’t do that there. Just simple misunderstandings about the legal framework that was never designed to work at a global level for a distribution platform like this.

Adam Bloomberg:

So, let’s talk about fair use. Can you define fair use for us?

Sam Rogers:

If I could define fair use for you, I would make a lot more money than I do. So, fair use is a legal concept that is basically a way of defining an exemption to infringement. And there’s a lot of different … It’s been defined really through case law more than anything, as far as what constitutes a fair use and how it is that works. So that’s slightly beyond what I would speak about in an expert witness case. But, just basically, if you think about a parody song as a classic example, I want to do a funny version of this very serious thing. That’s actually permitted in the US under fair use, in most circumstances, it’s generally a safer bet. That yes, if you’re Weird Al Yankovic, you can go, and you can do your funny version of the song that everybody knows. Good to point out, though, Weird Al Yankovic, someone who’s made his whole profession doing that. He always has an agreement as well with the artist that he’s parodying before he puts his parody out.

Adam Bloomberg:

That’s very smart. I was thinking Michael Jackson, Eat It, when you were giving me that example.

Sam Rogers:

Exactly.

Adam Bloomberg:

So, what constitutes transformative use?

Sam Rogers:

Transformative use is when something is building on the original; it’s superseding the original. And it goes all the way back to the 1840s when this concept was established, but keeping it in the realm of parody. So, you might remember there was a case, I think back in the 90s, with 2 Live Crew doing Pretty Woman. And it was one of those transformative use seminal cases, where they did their parody of the Roy Orbison song in their oh-so-tasteful way. And it came to court because they had actually asked for a license, and it was denied, and they went and did it anyway. And when they did, they made a bunch of money off of that song.

Sam Rogers:

And so, Roy Orbison wanted some of the money from that song, and it went to court because of that. I think it was considered a transformative use in the end, but it went back and forth several times. You can look up the case law; I think it was Acuff-Rose Music or something like that. But transformative use is actually one of the tests for fair use. It’s one of the things that determines if it is a fair use or not; again, this is a little beyond what I would generally talk about as an expert witness. But, I suffice to say, it’s a great thing for lawyers to argue about.

Adam Bloomberg:

Roy Orbison, what a great example; he’s such a great musician. I had heard of him first when Van Halen did Pretty Woman. I’m dating myself, but I’m such a fan of his with Traveling Wilburys, just the talent. You had mentioned with the National Geographic example, that five seconds of the cheetah running, with the Content ID, what is the typical time period that it prompts at? Is it 5, 10 seconds? Is it longer? Is it shorter? When does it get flagged?

Sam Rogers:

Much like with fair use or a transformative use, there’s no hard and fast rule about it. It’s something that the content owner can express, “No seconds of my cheetah running are ever okay.” They can do that. Just block it everywhere, and in general internet culture these days, I think there’s a six-second convention where people will allow clips of movies, meme length, basically. Meme length content, because it’s not denigrating the original, no one’s losing money because you have a picture of Jack Black making a funny face or something like that. A few seconds of that is just fine, and Jack Black’s okay with it. And so is his agent. But, if you try and upload the whole movie, well, now you’ve got a problem. “Okay, well, how about this scene? Well, how about this little mini part of the scene? Well, how about this longer part of the scene?” All of those are very unique cases that it’s really up to the content owner to decide, and YouTube’s Content ID tool is part of what helps them do that.

Thank you to Sam Rogers for speaking with us today, and a special thanks to our listeners.