Open Music Europe Policy & Governance: Feedback to the First Copyright Infrastructure Task Force

Open Music Europe Policy & Governance Feedback to CITF Copyright Infrastructure Task Force (Feedback Session) REPREX 2 December 2025 Oniline DOI: 10.5281/zenodo.17791740 Daniel Antal 2 Shared agenda Data sharing and exchange in the Slovak music data sharing space Local content regulation Help radio stations to comply with legal local content quotas. Data health checks Ensure that streaming services can recommend the music and pay out rightsholders Improve recommendations Music info centre sends authoritative repertoire data Improve statistics Satellite business register coordination for better surveys of music Integrate libraries Enable search in libraries and archives, locate printed music and records for public lending Printed music Increase the availability of printed music via webshops and library access OpenMusE | WP Presentation 3 1. Data enrichment Reprex consolidates the data Excel, CSV, SQL and other database (data file) formats of SOZA, Slovak Music Centre, Music Fund, and other organisations into a graph format. We assist curators to find and improve errors in their metadata. We ensure that the resulting data is more usable for rights management, heritage management, publishing. 2. Data dissemination We send the enriched and proprietary (confidential) data to the systems of SOZA, Slovak Music Centre, Music Fund or other participants. We sent the public data to the EU Open Data Portal (statistical data), to EOSC (data of scientific value), collections data (Europeana and ECCH), and biographical and repertoire data to Wikidata and Wikipedia. This way DSPs can use reliable data about Slovak music. OpenMusE | WP Presentation 4 Open Music Observatory Timeline 2020 2019 2026 Slovak Music Industry Report CEE Report Feasibility Study, Eu best practice, Yes!Delft AI Lab Digital Music Observator, JUMP Music Market Accelerator, SK Feasibility Study Various Use Cases, Grant Application Open Music Europe Open Music Observatory Open Music Observatory applications Full European federation AI Act CSRD Open Music Europe DGA, EIF Act Horizon Europe call Feasibility studies Triple Transition Music Moves Europe 2021 2022 2023 2024 2025 OpenMusE | WP Presentation 5 Shared Diagnosis: Fragmented, Incomplete, Non-Interoperable Metadata Aligned with detailed findings and vocabulary Implementation example Antal, D. (2025). A Green Paper on AI, Data Governance, and Metadata Policies for Europe's Music Ecosystem. Open Music Observatory. https://doi.org/10.5281/zenodo.17767905 Open Music Observatory: Federated Knowledge About Music Subsidiarity: sub-national, Estonian, Latvian, Hungarian, Finnish National very similar to the CITF setup Supranational: based mainly on open data Open Music Observatory: Federated Knowledge About Music Music: works, sound recordings (communicated to public and archival), print and manuscript sheets, lyrics. AV: music video Images: photos and creative work Text: journal articles, reports, books Data: microdata and statistically processed Shared Diagnosis: Fragmented, Incomplete, Non-Interoperable Metadata CITF: emphasises lack of interoperable, authoritative copyright data across member states Green Paper: stresses structural fragmentation, identifier gaps, misaligned workflows, legacy metadata problems OMO: identifies fragmentation across libraries, CMOs, labels, statistical authorities, music platforms, etc. European Interoperability Framework: Layers of Service Interoperability rability Framework: Layers of Service Interoperability Technical interoperability Semantic interoperability Organisational interoperability Legal interoperability Integrated service governance Rules of the data exchanges and use is harmonised to a level that negotiations and permits can be obtained fast to join the data. Organisations harmonise their internal workflows and jobs that use data to benefit the most from improving and enriching their own data with other sharing partner’s data. The jobs and workflows of the organisations in the data space share a vocabulary of meanings of the collected and shared data’s definitions and meanings. The data is translated to a standard graph format with shared annotation so that it can be easily exchanged, synchronized among partners. Shared Call for Open, Machine-Readable Identifiers All highlight the need for open identifiers across the lifecycle of works and recordings. Green Paper: emphasises alignment of ISRC, ISWC, DDEX, DCTERMS, RiC, DCAT, plus VIAF/ISNI authority reconciliation OMO: implements these identifiers in Wikibase; automates mapping; enforces canonical naming and PIDs across datasets CITF: insists on open identifiers as the backbone of EU copyright infrastructure and trustworthy AI usage scenarios Shared Recognition of GDPR–Attribution Tension All three acknowledge GDPR challenges in handling creators’ names while maintaining attribution. Green Paper: discusses the attribution vs. data-protection paradox directly OMO: treats GDPR-compatible authority control as core to the metadata model (VIAF/ISNI as privacy-minimising identifiers) OKM/CITF: emphasises pseudonymisation, opt-out structures, and data-protection-by-design for copyrighted content in AI workflows 1. Data enrichment Reprex consolidates the data Excel, CSV, SQL and other database (data file) formats of SOZA, Slovak Music Centre, Music Fund, and other organisations into a graph format. We assist curators to find and improve errors in their metadata. We ensure that the resulting data is more usable for rights management, heritage management, publishing. 2. Data dissemination We send the enriched and proprietary (confidential) data to the systems of SOZA, Slovak Music Centre, Music Fund or other participants. We sent the public data to the EU Open Data Portal (statistical data), to EOSC (data of scientific value), collections data (Europeana and ECCH), and biographical and repertoire data to Wikidata and Wikipedia. This way DSPs can use reliable data about Slovak music. OpenMusE | WP Presentation 19 AI Governance: All Agree AI Needs High-Quality, Rights-Aware Metadata CITF: analyses AI in predeployment and postdeployment phases, showing how rights can be violated or protected depending on metadata quality ) Green Paper: warns of AI misattribution, proposes “curative AI” for metadata repair and rights awareness workflows OMO: positions its data pipeline as a guardrail for AI, ensuring machinereadable rights, provenance, and identifiers AI Governance: All Agree AI Needs High-Quality, Rights-Aware Metadata Shared message: AI without trustworthy metadata endangers rights-holders; AI with trustworthy metadata enables new services. CITF: analyses AI in pre-deployment and post-deployment phases, showing how rights can be violated or protected depending on metadata quality Green Paper: warns of AI misattribution, proposes “curative AI” for metadata repair and rights awareness workflows OMO: positions its data pipeline as a guardrail for AI, ensuring machine-readable rights, provenance, and identifiers 22 Further information Reeprex B.V. Feasibility study for a European Music Observatory Key policy document guiding our work Identifies nearly 40 data gaps that burden both policy work and music business operations. These include data that are: -Scarce - the amount of available data is insufficient -Fragmented - data has to be collected from numerous different sources -Hidden - data might well exist, but cannot be found -Restricted - the data cannot be accessed due to technical or IP barriers -Unharmonised - the accessed data needs extensive and resource-intense pre-processing We have demonstrated that many of these gaps can be filled through open data reuse, cross-stakeholder data sharing, improved surveying, harmonisation, and other novel methods. Starting point 1. Why online platforms do not recommend music from Slovakia in Slovakia for Slovak people? 2. How can we avoid that small repertoires lose their original market, their only market? Findings 1. Small repertoires have very low-quality data representation 2. Because of low income, data improvement and documentation must not be expensive 3. Public-Private Partnership is needed to solve data problems OpenMusE | WP Presentation 25 The opposite of success It is often hard to say what would be a perfect metadata, or what we want to do with the data. We started with the nightmare scenario: that a new release ends of on Forgotify, a funny service that plays songs which had never been played on Spotify. We realised that many songs, often locally well-know records, end up here with wrong data, because recommender systems place them on the wrong shelf and recommend them to an inappropriate audience. Review of data catalogues, faceted search, API, chatbot Reprex B.V. One Song, Two Paths of Use Different licences govern what you can copy —and what you can play. # --- Commercially released song ---------------------------------------------- wd:Q4745 a wikibase:Item ; rdfs:label "Tšitšōrlinki, tšitšōrlinki (commercial release)"@en ; schema:description "Commercially released version of the traditional Livonian folk song, performed by Hilda Grīva."@en ; p:P410 s:Q4745-spotify-access . # access point → Spotify link (statement node) # --- Statement node for the access point ------------------------------------- s:Q4745-spotify-access a wikibase:Statement ; ps:P410 <https://open.spotify.com/track/123456789abcdef> ; # Spotify player URL pq:P483 wd:Q5989 . # qualifier: has policy → Listen on Spotify # --- The policy item ---------------------------------------------------------- wd:Q5989 a wb:UsePolicy ; rdfs:label "Listen on Spotify"@en ; schema:description "You may listen to this recording via the Spotify web or mobile player, free or by subscription, according to Spotify’s End User License."@en ; dct:license <https://www.spotify.com/legal/end-user-agreement/> . Broad Interoperability for Diversity: From Archive To Spotify Our data model supports only “patterns” of important standard library, archive, museum, rights management conceptual models; functionality is not optimised for libraries but for cross-institutional use 1 2 3 Full description for curators, promoters, researchers •Using the familiar Wikipedia/Wikidata GUI Unrestricted access on legal streaming services for all •On Spotify & other platforms, educational, personal use without copying rights. Limited (non-commercial) access for research •Europeanan and the EU’s new Culture Cloud 1 2 3 Find Our Results OpenMusE | WP Presentation 35 Papers, presentations, datasets, social media Website: https://www.openmuse.eu/ LinkedIn: https://www.linkedin.com/company/openmusiceurope/ X: https://x.com/OpenMusicEurope Instagram: https://www.instagram.com/openmusiceurope/ Repository: https://zenodo.org/communities/music_observatory/ Thank you for your attention! OpenMusE | WP Presentation 36 Get in touch for replication advice, questions, comments Dániel Antal www.linkedin.com/in/antaldaniel Reprex reprex.nl/contact Tomáš Mikš https://www.linkedin.com/in/tom%C3%A1%C5%A1mik%C5%A1-a9137441/ SOZA https://www.soza.sk/page/contact