From a Registry to a Registry Ecosystem
Today, we publish a new strategy paper on how Europe’s emerging copyright infrastructure should be developed. Authored by Paul Keller and Sebastian Posth – and developed as part of the CommonsDB pilot – the paper distils 18 months of building and operating a working public registry of the public domain into a proposal for how a system of copyright registries should be organised and the role that public institutions should play in it.
For most of its history, European copyright law has functioned without formal registration. Protection arises automatically, and there is no public record of what is protected, by whom, or under what conditions. This was appropriate for as long as copyright governed a slow-moving world of professional intermediaries. But automated systems now make decisions relevant to copyright about billions of works, such as filtering uploads, assembling training data for artificial intelligence, and resolving reuse permissions. Those systems depend on reliable, machine-readable rights information. But currently there is no public infrastructure that can provide it at scale.
Much of the urgency stems from generative AI. European law gives creators the right to opt their works out of AI training, but there is still no reliable way to express an opt-out so that it reaches the AI companies required to respect it – and no reliable way for those companies to find the opt-outs they must comply with.
We first analysed this problem, and put a registry-based answer on the table, in our 2024 policy brief Considerations for implementing rightholder opt-outs by AI model developers. Registries offer similar benefits wherever works need to be reliably identified: as being in the public domain, as openly licensed, or recognised as orphan or out-of-commerce works. Registries have moved from the margins to the centre of copyright policy.
The open question: how should a registry ecosystem be organised?
This year, two studies commissioned by EU institutions put registries firmly on the European agenda. A study for the EUIPO mapped existing copyright databases and proposed a federated layer to connect them rather than replace them with a central database. A feasibility study for the European Commission concluded that an EU-level registry of AI training opt-outs is feasible and recommended an architecture for it.
However, both studies left open the question of how this infrastructure should be governed, and how an ecosystem of registries should be organised, including what role public institutions should play within it.
What does CommonsDB show?
Our strategy paper provides answers based on experience. With CommonsDB, we have built and operated a working public registry of rights information for works in the public domain and those with open licences. With over 6.6 million verified declarations, it is the largest known public deployment of a registry built on the ISCC, an open ISO-standard identifier derived from the content of a work itself, to date. Five conclusions can be drawn:
- Purpose-built registries work. The CommonsDB pilot demonstrates that such registries are technically and operationally viable at scale.
- Registries belong close to those who make the records. Rights information is best collected and governed by the organisations and communities that know the works: they can verify claims, correct errors, and answer for their records.
- This means there will be many registries. Different sectors and purposes need different registries, and no single central database can absorb them all.
- Federation makes many registries work as one. CommonsDB has shown that independent registries can publish and synchronise their records over a common protocol, so that anyone can query them as if they were a single registry.
- This landscape needs a public anchor. In an ecosystem of many independent, federating registries, the task for public institutions is to set the standards, certify the registries that meet them, and run an aggregation layer that provides one authoritative view across all of them. The paper proposes the European Union Intellectual Property Office (EUIPO), which already operates the EU’s two existing copyright databases, as the natural home for these public functions.
What is at stake reaches beyond copyright administration: machine-readable knowledge about European works should be held as open public infrastructure, rather than accumulate exclusively in the data portfolios of the largest platform and AI companies.
Our new strategy paper develops these conclusions in full. It is accompanied by a legal annex, written by our partners at the Institute for Information Law (IViR), which examines what the EUIPO can already do under its existing mandate and where new legislation would be needed.
Open Future has led the CommonsDB initiative, which is funded by the European Commission. In collaboration with Liccium, the Europeana Foundation, Wikimedia Sverige and the Institute for Information Law UIViR), we have brought the vision of a prototype registry of public domain and openly licensed works to life.
The CommonsDB project was co-funded by the European Union under the 2023 work programme on the financing of Pilot Projects and Preparatory Actions in the field of Communications Networks, Content and Technology.
