The third stakeholder dialogue took place in Brussels on Monday 25th November.
It has been webstreamed here. Time of appearance are referenced below.
YouTube (from 10:18:.. H ... to 10:40:...H )
YouTube was the only plateform to make a presentation. It was informative as, beyond the tools described therein, they disclosed their line of thoughts on the application of Art.17 specifically when it comes to still images. Their presentation drew many questions from the stakeholders in the room and from the European Commission. For the purpose of this report I will focus on the questions with respect to the use of image from 10:52: ... H 11:03:48 H
They presented their tools: 1. Takedown Webform 2. Copyright Matchtool 3. Content ID And pointed to the challenges: finding same files rather than similar files in a database containing 600 years of reference files, cross-referencing rights
Carola Streul asked whether the Content ID was available for images, YouTube's answer was no. And they clarified why: YouTube is a video platform, not an image platform.
Their business model is to monetize audio-video content, as a consequence thereof they develop copyright management tools to identify and distribute revenues to content owners. Since they have no interest in images, however, there is no commercial incentive for them to develop a content ID for images.
In their view, this "why" and "how" platforms are build should be considered upon applying Art.17. , when determining "best efforts" "in the light of ther principle of proportionlality".
In other terms, they read the article below that it would not be "proportionate" to apply Art.17 to images when the business model of the platform is not about images.
Art.17-5 states the following:
17-5-
Sylvie Fodor, representing CEPIC, asked them to clarify their point of view at 11:25: ...H. They then explained the technical difficulties in identifying images.
I used my question time to remind that, unlike they were saying, images are not "incidental" in YouTube videos: very often YouTube videos are made up of images in the forefront, with some music in the background. The point is that if images are not "incidental" in YouTube videos, it might not be "disproportionate" to require from them to develop a Content ID or any other means to make sure that infringed images do not say on their platforms or to include the use of images in potential licenses.
Videntifier (from 12:28 H to 12:42 H) and PEX at 12:04:07 H
The Videntifier presentation is worth a special mention as it concerns image recognition technology and is very clear. It is attached.
Videntifier is an image recognition technology using fingerprint working in combination with a crawler. It is used by law enforcement authorities and, newly, by the ADAGP or Facebook since early 2019. Questions relate to how the technology works.
PEX have an audio and video search engine and are working with rights holders. Interresting is that they have developed what they call an "attribution engine" to work with platform and identify rights. They propose this engine for free to platforms against the share of the revenues of advertising. (At the moment no platform has shown interest in working with PEX.)
Both technology vendors insited that their technology is affordable by small platforms, not just by larger ones.
EVA (from 14:50: .. H ... to 15:02: ...H followed by questions until 15:30: ... H)
EVA's president, Vincent Van Den Eijnde, made the presentation.
(NB: Vincent is the CEO of PictoRights, CMO in the Netherlands. They were part for some time of NL Images.)
The two shots below illustrate their master plan. The future envisaged is based on what is existing at present. Extended Collective Licensing for instance is already in use in Sweden and works well.
CMOs for visual arts see themselves at the center of a virtuous eco-system for licensing images with platforms. Extended Collective Licensing is the magic formula to solve all issues. Art.12 of the EU Copyright Directive allows for extension of ECL to the rest of the EU. It is a flexible solution allowing for opt-out.
They say they represent photographers and work with picture agencies. Carola Streul specified that so far there were only two licensing deals with platforms: the ADAGP has a deal with Daily Motion and another one with Google concerning images in YouTube.
As a conclusion the EC clarified the procedure of the Stakeholder Dialogue.
Phase 1 - meetings 1 + 2 - provided sector specific overviews
Phase 2 - meetings 3 + 4 - provided specific market practices
Phase 3 - meetings 5 + 6 will provide forward looking practices
Future meetings:
The meeting of the 16th December will have the same format.
Stakeholders who are not able to make a presentation can send papers that will be shared with everybody.
The "save the date" for the fifth and sixth meetings are 16.January and 10.February.