The following are the outputs of the captioning taken during an IGF virtual intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.
>> ELENA TAIROVA: Okay. That was beautiful. Can I now have an opportunity to share the slides because it's 3:35.
Okay. I believe I can share the screen. Give me one second, and we will start.
>> DANIEL MARICIC: While Elena is doing that, let's, instead of silence, talk.
Welcome, everybody. If you wouldn't mind turning on the cameras, let's see you all and let on if that's already with all of you, I will take a screenshot of everybody just sharing on social media. Of course, if you don't want to be on the screenshot, turn off the camera, and that's it.
Yeah. So some of you don't know me. Some of you do know me.
My name is Daniel, and I'm a cofounder and co‑organizer of this talk.
Together we have with us, in order of my screen from left to right, we have Robert and Sylvie, who will be presenting with me. I will be at the end.
Elena you know and Ali, they are team members.
Hi, George, nice to see you from the USA.
Adriano is part of the development team.
Then we have two IGF, welcome.
And captioner and Dominic, welcome, as well.
So, yeah, feel free to take up the mic and discuss. Let's not make it like me talking.
Let's start to my right.
Robert, would you like to say something? Or let's just talk a little bit about stuff. Unmute and introduce yourself, if you want to.
>> ROBERT HERIAN: Good afternoon, Daniel.
Good afternoon, everyone.
Yeah, I'm Robert Herian. I'm presently based at the Law School at the (?) University here in the UK. My interests are based in law and technology. It's kind of across law and property law. That's why I'm here. I'm looking forward to some interesting discussions today.
I will hand it over to Sylvie.
>> SYLVIE FODOR: Good afternoon. Yeah, I will make a presentation later when we can share the screen.
I'm Sylvie Fodor. I'm the director of a trade organization, which means we present the federation. We (?) Visual content producers, is the large meaning.
I'm talking here from Berlin. I'm at the home office. I'm talking from my home, but we have an international membership everywhere in Europe anyway. We've been very active in Brussels.
I will share a bit more when it's my time to speak.
If we can share the screen! I don't know. Maybe we cannot share the screen, and we can just say something. I can share my own screen, maybe. I don't know if it's a problem. I'm just trying to help.
>> ELENA TAIROVA: Thank you for your patience. It's still struggling. I'm still struggling to share for some reason. I don't think I have the rights.
>> ROBERT HERIAN: Excuse me. Apparently, I have it.
>> ELENA TAIROVA: Okay. It's beautiful, Daniel. Can you please ‑‑
>> DANIEL MARICIC: Sure. (Laughter). I didn't notice it. I'm sorry.
So what ‑‑ I go with the center.
>> ELENA TAIROVA: If you can do that, yes. That will save us.
>> ELENA TAIROVA: Perfect! Here we are.
Okay. So, basically, you guys, I guess, all know since you joined the agenda that we're going to be talking about the remaining 30 minutes about Digital creative market: How can we protect users and ensure freedom in light of the new EU Copyright Directive?
We'll be talking about technical uses and information and emerging technology. We'll finish the discussion and then open the floor to Q&A.
So let's go right to it.
First, Sylvie will give us a perspective of the creative industry. Sylvie, please take the floor.
>> SYLVIE FODOR: Thank you. Thank you, yeah. Okay.
So thank you. I'm very happy, very honored to be here and to be talking about this important subject on behalf of our varied membership.
If you could, go to the next slide, please.
So that's a presentation of CEPIC. I thought I would make a presentation. I'm not sure everyone is familiar what trade organizations do and what CEPIC does. So I said we would present visual content producer. This means a picture agencies but also photographs, videographers. This is not as detailed as you will see in my presentation.
Everyone is in Europe mostly, but we are registering ‑‑ we just registered as an international organization.
So what do we do? We share information between members, and we also fight for the interest of our members. When I mean fight for the interest, I mean mainly that we fight for copyright, for copyright online.
This is why I'm here on this talk. And this brings us to our next slide, please. Yes.
So that's why we've been very active in the debates in the making of the (?) Directive, and we focused all of our efforts on Article 17. Corporation legislation has been one means among others, to solve the value gap or the value block. What is that? The value gap is the name given to the disbalance that you have online with value distribution, value of content distributed by users, shared by users, with online platforms such as Facebook or Google, making a lot of money thanks to this creative content, and creators often making very little money, if any money at all. That's why we call it the value gap or value block.
We fought with this as a tool to solve the value gap.
Now, the thing is ‑‑ and this is a conflict here ‑‑ copyright and, of course, Article 17 is very often (?) It means their free speech is being impeached. I don't think ‑‑ or members don't think this is the case. I don't think this is the case. I think it's a contra decision ‑‑ it's contradiction. Creativity is a freedom of speech.
So that's for the theory. Now, how do we solve this question in the real world. How can Article 17 support creativity and freedom of speech?
Technology is key. We've always said technology is key. We have several companies that help us launch consultation business providing information, and they're looking at the market.
So technology is the key. It's at least one very important part of the solution.
On this slide, I made a difference between technology existing before Article 17 and then after Article 17 came into being.
Then technology that is compatible with Article 7 and then one that enforces Article 17.
It's quite obvious that you need a market for this. So before Article 17 came into being, we have to admit the market did not exist for it; right?
At the moment, we're in a period of transition. We're in a period of transition, 2019, 2021. We have things in progress. We have several works on technology that will enforce Article 17 in full. However, what we had is we already had technology that could support this Article 17. That's what we had before. That's what was part of our consult at the EU. Now you've seen my quote with the translation of the report that was made at the French government on this implementation of Article 17, that recognition tools play a role in licensed agreements by (?) The content and dissemination of work on the Internet.
It's intended to extend and broaden the dynamic, which is favorable to creation and has been initiated on certain platforms. Hence the increased importance of content recognition tools. That's where we are. We have technology compatible with Article 17.
I will give one example here for the presentation (?) In 2018. We that it's compatible with Article 17. It covers a number of aspects of Article 17. Then we still need and are still working on the technology that will enforce properly this article.
Yeah. And that's it, I guess. That was my message for today.
Thank you very much.
That's my last slide.
>> ELENA TAIROVA: Perfect. Thank you very much, Sylvie. You were perfect on time.
Now we give the mic and the floor to Robert, which gives the perspective ‑‑ legal perspective on the current work in the field of regulation.
>> ELENA TAIROVA: Robert is senior lecturer at a pioneer in terms of (?).
Robert, it's all yours.
>> ROBERT HERIAN: Thanks very much, Elena.
So I just have the single slide. So you can all entertain yourself by reading that statement through my very short presentation.
If starting point is copyright law that contains various principles and variables. Copyright law provides an automatic right when work is created and it's the owner of that right, in a majority of jurisdictions, will need to identify and report infringements.
So any suggestion of an absence of law, really, in this regard is incorrect, whether we're talking about civil or common law jurisdictions. But just because laws are in place do not mean that they are enforced or, indeed, enforceable in any realistic way.
We have a gap similar to Sylvie's one. It's one that continues to breed uncertainty and a lack of confidence.
As the 2004EU directive on international property right, (?) Is low or lost when a creator risks not being able to derive or have the opportunity or potential to derive a legitimate profit from their interventional creation.
We know (?) Can facilitate wide economic benefits for communities and for countries.
So enforcement rather than a need to evidence a prima facie right. It perhaps makes things more easy or equally detectible.
Article 17 of the copyright directive requires online content sharing service providers, or OCSSPs have to add (?) Before adding content to a platform. And they will be liable for the content they share unless they obtain an authorization, two, block unauthorized content and ensure it remains blocked, and, third, remove unauthorized content once notified.
The question is whether such measures can enforce weaknesses when breaches occur.
So an event yule creator cannot a‑‑ eventual creator cannot (?) Their know‑how.
In the UK and the U.S. and EU countries like France and the Netherlands, there's content providers and other companies that facilitate access to material that infringes copyright online and the demand side, so that's from individual subscribers. Though the approaches do differ, there are universal aspects to the things put in place regarding copyright infringement.
This is increased warning and fines and policies for taking down content in case us of supply site infringements as well as the follow‑the‑money approaches that target stakeholders involved in copyright infringement, such as those that monetize content.
Now, (?) Against copyright infringement is (?) Although effectiveness is really debatable.
South Korea is the (?) Including extrajudicial actors who have a responsibility for reporting infringements.
Dealing with copyright issues is labor intensive. It's a form of laboring in which the public, for example, are not compensated. Although.
>> > We may argue that there's a fair obligation to justify a fair space for all which outweigh s (?) An individual might expect.
The obvious response seems (?) In relation to (?) There are semantic systems that are able to create provisions and deploy them as effectively as extrajudicial and means presently to hand. But I do want to stress that this is not an answer rooted in a technological workaround of the law as we find it offline. It's really about technology using laws that, in many cases, we already have. So that's technology finding itself in the law, therefore, not the other way around.
So back to you, Elena.
>> ELENA TAIROVA: Thank you so much, Robert. I heard from the organizers, they kindly gave us five extra minutes since we were unable to start on time.
Without further ado, Daniel, please continue with the technical perspective on the issue.
>> DANIEL MARICIC: Okay. Let's not lose more time.
I'm going to run through this, Robert.
So to kind of connect what I'm going to talk about, which is the technical part, is real‑world copyright is very hard to verify in the digital world because you need identifiers.
Today, we mostly see technical solutions focused on preventing approach like filters for blocking or other ones and non‑working technology that is used to actually enable digital copyrights and create them as they should be created in a digital world.
This slide, one of the biggest challenges in the multimedia domain is to identify properly and the standard does not enforce creation of unique identifiers. Due to that reason, we need to create another approach aside from the metadata standard. We need three identifiers, at least. Multiple identifiers are going to be a combo of identifiers.
The image you see here is a working prototype of a multigenerational workflow which creates two cryptographic identifiers and one perceptual feature‑based identifier that's widely used in a reverse image search and in an AI‑model training.
We saw that the common identifiers are important in generic creation of a digital copyright. Now we'll talk about important identifiers which enable the connection between the real‑world assets and the user identity in the digital world.
On this screen, it's an output of a program, eight images with three taken with a professional camera and four with a mobile phone.
The camera model plus the serial number is the unique identifier of such a device, which, unfortunately, does not have information enough for the mobile phones, which is the majority, which are also the devices used the majority of the time, most of the time in creating the images.
At the end, the original document ID, as defined in the standard, says this is the mother identifier with which all comes, and it's possible to track them to the older renditions and modifications which, unfortunately, mobile phones do not provide.
And that's not enough of violation of the metadata standard. The mobile phones are actually providing some identifiers in a software definitions which is not meant to be used like that.
The software should be there to identify which software has been used to modify, not create. So this is the real‑world asset problem.
This slide is about connecting the previous two slides. There's a saying. If you have access to a private key, you have access to Bitcoin or a (?) Statement.
This version consists of the account ID ‑‑ (distorted) ‑‑ that acts as real‑world classical signatures and it contains the proportion amount (?) And additional research is proof of existence of the real‑world asset from the previous slide.
On the other hand, all of this is possible to be ‑‑ the only way to make this transparent and audible is through this thing called role ID which is actually a clean, transparent workflow of executing from data to the actual proofs, which are bundled into claims, which are bundled together into the final state.
>> ELENA TAIROVA: Thank you, Daniel.
>> DANIEL MARICIC: We have five minutes for Q&A.
>> ELENA TAIROVA: We have 10 extra minutes.
If anyone in the audience has anything to add, any questions, particular questions to our speakers, now is the time.
Unfortunately, I believe most of the audience is on a YouTube channel, watching the YouTube stream because of some technical issues.
So people who are here, you probably need to ‑‑
>> DANIEL MARICIC: Should we pick anybody just randomly. Come on, don't be afraid, people. It's all right.
>> ELENA TAIROVA: Still, if you have something to add, you have extra time.
>> ROBERT HERIAN: Yeah. I mean, I have a question to put back to Daniel, really, which is I suppose the unpacking of transparency, in your slide there, you described transparency as you were passing the (?) For a series of letters and numbers.
>> DANIEL MARICIC: Yes.
>> ROBERT HERIAN: In other words, it's not what most people would consider an ordinary transparency.
How are we more inclusive?
>> DANIEL MARICIC: We're from different domains. How do you define inclusive? It could mean ‑‑
>> ROBERT HERIAN: Inclusive in the sense of those who may not understand a series of figures, letters, or numbers as transparent.
>> DANIEL MARICIC: Right. Well, this is not meant to be read by end user. This will be ‑‑ how we call it? ‑‑ beautify ed by the UI or applications. This is what I showed in the last slide, it's the output from the ‑‑ it's a structure of the statement. All the values are just actual values, and since there are strings and letters. They are actual links to other things. If you would have a UI or an application that you would say, Hey, I would like to see my statement or my copyright or this image. Does it have a copyright?
It would say yes or no. If it has a record, it would have it. If it can't find a record, it would know. If it find a record, then, okay, what are the details.
So it can be shaped in a date or however users decide to show it.
>> ROBERT HERIAN: Just to follow‑up, I'm quite interested in this. It may seem perhaps somewhat irrelevant, but I'm wondering how important it really is. If we're going to talk about transparency seriously, when does transparency really begin in the process? If it only begins at the point of, as you say, kind of an easily interpret end‑user interface, is that too late?
(Captioner has no audio)
>> DANIEL MARICIC: Sorry for that. Transparency ends ‑‑ the end cycle of transparency. The beginning of transparency is ‑‑ if you think of it, my left hand is data, and then in my first slide, I talked about workflow and the awesome thing that goes from left to right. And there's an image of the universe on the right‑hand side. That's where the transparency starts.
So on that right‑hand side where you start creating all of the proofs ‑‑ we call them proofs, but they are essentially identifiers that act as a proof.
So if you want to say is this image equal as this one, there are two ways to do it, either cryptographically, which you aim for 100% uniqueness, or local (?) Which provides similarity on a feature base. For example, if I were to look at all of our images here, I would crop them and run them through the perceptual (?) We would have 90‑plus similarities because we all have a head and roughly around the center of the screen.
So the higher the perceptual hash is similar to another perceptual hash, it's more similar ‑‑ the images are more similar. Basically, 100% is an image.
Also, if you say you have three of the identifiers and two says nothing, it's not equal, they're cryptographic in nature, there's a change and you can say, Okay. I give more weight to the perceptual hash, but it's subjective. The idea, how we think of it, we do not enforce the weight. Actually, its weight is on you. You would either get on the court and defend that and say, yes, it's more (?) Than crypto‑‑ it's very easy to change.
So each step of the way, to the end of the statement, it's transparent. You can see what exactly has been down, how it's been down. What you cannot see is the data. The data is not saved. The data never leaves. The user ‑‑ I don't know ‑‑ either UI or mobile phone or environment. It's never stolen.
We're creating like we never touched it.
I hope that answered your question.
>> ELENA TAIROVA: Thank you, Daniel.
Right. So a few minutes, if anyone in the audience has a question.
Otherwise, I suggest if nobody raises their hand, I suggest Daniel, Robert, Sylvie, please, a closing statement from you on the topic of our discussion.
(Captioner will disconnect in 60 seconds)
>> DANIEL MARICIC: Nobody has anything to add.
>> SYLVIE FODOR: It would have been interesting to know ‑‑ if I could say something ‑‑ I know we don't have so much time, unfortunately, but the motivation of the people who participated in this session ‑‑ as a reason I'm saying this is usually I do this kind of presentation in front of specialists, people who are really dealing with the subject on a day‑to‑day basis. I don't know if anybody wants to say something, but it may be ‑‑ maybe after the presentation ‑‑ if we could know, as a motivation ‑‑ yeah, the motivation. It would be interesting, from my perspective, as a presenter.
>> ELENA TAIROVA: The platform allows you to send a message to the organizers, and you can use the #L230 which is the hashtag on social media and we can follow up after the call and continue the discussion if you have anything to add.
>> DANIEL MARICIC: That's it? No questions? Nada? Nada?
George. George is asking something.
You can speak.
>> George: I've got power. For God's sake.
I'm calling if Los Angeles.
Do you see any litigation with contracts in the blockchain?
>> ROBERT HERIAN: Hi, George. Thanks for the question. No, there has not been a great deal. There's some interesting sort of movements around continuing to look at sort of definitions of cryptocurrencies in terms of property which tide back to the blockchain but not necessarily ‑‑ I mean by "contracts," are you talking about smart contracts?
>> George: Yeah. Sorry.
>> ROBERT HERIAN: Not really, no. Because we just completed ‑‑ well, the Law Commission just completed it, and I was a part of it, (?) Of the present law of contract in relation to smart contracts and reached the conclusion that, actually, the existing law is capable of catching any issues that may arise in relation to smart contracts.
I would imagine there may now be sort of litigation that follows in the wake of that because there will be people essentially testing how rigorous contracts ‑‑ or traditional contract law and theory is in relation to what smart contracts can do.
Now that that's been set out, I wouldn't be surprised if we start seeing the more deployment of smart contracts and, therefore, more failures of certain situations which were needed to be tested in court.
So, yes, they're definitely on the horizon, but nothing significant at the moment.
>> George: Thank you.
Hey, Daniel, I have a question ‑‑ or I had one. It just escaped me.
Is there monetization?
>> ELENA TAIROVA: I see we're not kicked out, even though we're at time.
But just in case the recording stops, I would like to thank everyone for joining us today.
(Captioner is disconnecting)