Clear standards on "open" are needed to introduce consistency to practices and legal certainty around reuse
The disharmonized use of “open” curtails usability and introduces uncertainty around reuse permissions. Its (mis)use to describe digital access to collections and/or the release of proprietary or low quality data is an adoption of “open” in name only. Areas where greater clarity and standardization will benefit the open GLAM ecosystem include:
“Open” must communicate a reliable standard for reuse purposes. Open access should carry a meaning that provides clarity to the user and reduces risk in engagement. Considering the complicated state of copyright and other rights that may legitimately restrict access or reuse, it is crucial to clearly distinguish which collections are open access and which are not. Non-open access is both justified and necessary in physical and digital environments.
“Open” should include qualitative aspects of access for reuse purposes. Varying technical standards that are used for resolution, format, or metadata can compromise reuse of open-compliant materials.
Using tools that already exist. The machine-readable rights statements already in use by the GLAM sector are designed to increase legal certainty around cross-border reuse.
Creative Commons tools. Both Creative Commons tools are “open.” These include the Public Domain Mark and the CC0 1.0 Universal Public Domain Dedication. Creative Commons recommends the latter tool for the release of non-original reproduction media because it provides greater legal certainty for users.
Creative Commons licenses. Not all Creative Commons licenses are “open” according to international standards. Open CC licenses include CC BY and CC BY-SA. These licenses rely on the existence of copyright, so should not be applied to non-original materials.
Labels as distinguished from licenses. Unlike licenses, labels can be applied by anyone when digitizing both in-copyright and public domain collections and communicate the status of the highest level of rights present in a digital object.
Cultural Permissions. There are frameworks specific to the management of traditional knowledge and traditional cultural expressions that sometimes map onto IPR frameworks or sit outside of them entirely.
Layers of rights and misapplication. Conflicting statements and mislabeling can result from conflating the IPR status of distinct layers in an image, a lack of information around the source work or another layer, or misunderstandings around the appropriate tool, license, or label to apply.
Building consensus around using licenses (or not). Greater consensus is needed on whether rights arise, as well as whether they should be claimed for various media produced during digital initiatives.
Building consensus around respecting tools, licenses, and labels. The wider GLAM ecosystem, particularly users, are also key to improving practices and understandings around the use of tools, licenses, and labels in digital environments.
Clear open GLAM standards on “open access” and how to engage with it would help introduce consistency to practices and legal certainty to public reuse. Otherwise, “open” in relation to “access” will continue to reflect subjective preferences.
The OpenGLAM initiative should adopt a clear standard aligned with international definitions, which condition “open” upon allowing commercial reuse. These include, for example, the Budapest Open Access Initiative,1 the Bethesda Statement on Open Access Publishing,2 the Berlin Declaration on Open Access to Knowledge in the Science and Humanities,3 the Open Knowledge Foundation’s definition,4 and the standards used by the Open Access Scholarly Publishers Association.5 This means only materials published under public domain, CC0, CC BY, CC BY-SA, and similar statements (e.g., “No known copyright restrictions”) are open-compliant under international standards.
Doing so will increase reuse certainty, benefiting users and GLAMs in three immediate ways:
It will signal “open” materials can be used for any purpose, enabling their integration into public platforms that condition uploads upon allowing commercial reuse.
It will bring greater clarity to the sector around what activity and materials qualify as open GLAM, rather referring to activities that extend only digital access and limit reuse of public domain collections.
It will help validate various non-open activities and materials that cannot be “open” due to legitimate legal or ethical reasons.
“Non-open GLAM” is not a dirty phrase. It just means institutions may be unable to release their collections under open frameworks, like the reproductions of in-copyright collections or sensitive materials. Yet, this does not mean that non-open GLAMs (or non-open collections) have nothing to contribute to the open GLAM landscape. Most collections data or metadata can be released under open frameworks to allow for their integration with other data. This is also important in order to balance the risk of in-copyright collections being excluded from or obscuring the new connections that might be made.6
The OpenGLAM initiative could help clarify reuse limitations to users and defend non-open GLAMs, by helping draw a distinction between a non-open GLAM (i.e., unable to participate for legal reasons) and a NOpenGLAM (i.e., able to participate, but refrains). There has been a demonstrated need to clarify reuse limitations for users’ benefits.7 Doing so may help the public better understand, appreciate, and respect distinctions around in-copyright versus public domain, and particularly with respect to the underlying work’s IPR or other rights and appropriate reuse during engagement with digital collections.
It is equally important to consider qualitative aspects of open. The varying technical standards that are used for resolution, format, or metadata can compromise reuse of open-compliant materials. Even if no rights are claimed (or if open licenses are applied), GLAMs might release images that are low resolution, embedded with messy or nonexistent metadata, or lacking machine-readable rights statements.8 These variables pose difficulties to meaningful reuse, disrupt interoperability, and render some collections unusable for purposes other than viewing online.
These technical aspects demonstrate the need for further discussion around baseline standards for resolution, interoperability, machine-readable formats, and other technical standards to facilitate reuse of open GLAM data. This point is revisited in Technical Standards.
There are at least four benefits to using existing machine-readable rights statements and standardizing approaches among (and within) GLAMs.
The free text typically contained in copyright metadata fields and managed via digital asset management system is not machine-readable, nor can it be read without technical extraction. By contrast, machine-readable statements are searchable and can be read by human and machine users, integrated into data aggregators, and filtered by major search platforms.
Standardized licenses and labels are both well-known and multilingual, which allows downstream and cross-border users to read the statement terms in native languages.
In the absence of IPR harmonization on various reproduction media types, standardized and machine-readable licenses and labels provide legal interoperability, which should be maximized for cross-border online environments.
These initiatives are stewarded by independent groups of people who monitor international legal changes and provide necessary updates. This cannot be said of manually embedded data fields and statements, for which the logic and reasoning of bespoke statements may be provided by staff who will inevitably leave the GLAM.
With this in mind, we’ll examine what machine-readable rights statements are available, and how they may be used (or misused).
There has been a wider debate around which tool is more appropriate for the release of non-original materials. Some advocate for use of the Public Domain Mark, because it confirms the legal status of the work in a way CC0 does not. The text of the CC0 tool suggests that rights arise, but that the rights holder has dedicated the material to the public domain. On principle, the use of the CC0 Public Domain Dedication is therefore inappropriate: the “rights holder” cannot dedicate what they do not legally hold.
However, there is an important point to highlight here. Just as Copyright discussed various parameters around copyright protection and its territorial differences, there are similar differences arising with the public domain. National standards for originality or related rights may render materials as public domain in one jurisdiction, while it would be subject to protection in the next.11 For some users, this poses significant risk to cross-border reuse online. The CC0 1.0 Universal Public Domain Dedication recognizes this uncertainty and ensures users that if any IPR or other rights would be recognized in the jurisdiction of use, the rights holder has waived them or agrees to a fallback license (hence the “Universal”). Creative Commons recommends CC0 for reproduction media of public domain works for this reason.12
To be clear, CC tools may be applied only when no rights arise (or have been waived) in any part of the composite media.13 Applying the Public Domain Mark or CC0 1.0 Universal Public Domain Dedication tool to non-original reproduction media of an in-copyright work is neither appropriate nor legal (unless the rights holder of the underlying work has agreed to release the rights and the tool’s application). Alternatives for this are discussed below, namely labels.
Licenses are appropriate only when IPR arises, and can be applied by only the rights holder.16 Accordingly, if the underlying work is in-copyright, and the GLAM does not own those rights, CC licenses are inappropriate (see the discussion on labels instead). This means GLAMs can apply Creative Commons licenses to:
Non-original reproduction media for which the underlying work is in-copyright and the GLAM holds those rights; and
GLAM materials that satisfy national standards for copyright protection.
It is worth noting that while the CC BY-ND (No-Derivatives)17 license permits commercial use, it is not “open” according to the Budapest Open Access Initiative and its 2012 recommendations.18 This is because the ND restriction will prevent modification, translations, adaptations, and other useful activities, including how subsequent research using the materials might be shared.19
The remaining CC licenses are not open under international standards due to their restriction on commercial use. These include: CC BY-NC,20 CC BY-NC-SA,21 and CC BY-NC-ND.22 Many legitimate reasons lead GLAMs to choose closed licenses for GLAM-generated intellectual property. Guidance for that decision is discussed in below, and in GLAM-Generated IP. It should be noted here that the application of closed licenses prevents engagement with external platforms that require open-compliant licenses.
Labels can be applied by anyone when digitizing both in-copyright and public domain collections, rather than only the rights holder.23
Labels are designed to communicate the status of the highest level of rights present in a digital object. Those most commonly used include the twelve standardized statements designed by the Rights Statements initiative.24 As applied to non-original reproduction media, Rights Statements are useful to reflect the IPR status of the underlying work. For example, if a book is in-copyright, the digital surrogate of the book will be labeled “In Copyright.” No rights arise in the digital surrogate, but reuse is restricted by the in-copyright status of the underlying work. By contrast, if the book is in the public domain, the digital surrogate of the book will be labeled “No Copyright.” Neither the digital surrogate nor the underlying work presents any copyright reuse restrictions.
Three Rights Statements align with international standards for open. These include: No Copyright, No Copyright – United States,25 and No Known Copyright.26 Because the label applies to the highest layer of rights that might arise, use of the label communicates there are no IPR claims in the digital surrogates.
Here, a distinction should be made on the difference between the legal statuses conveyed by licenses and labels. All Creative Commons licenses and tools communicate a clear legal status of the work. The user can be certain their reuse is legally compliant, so long as it falls within the license’s parameters (and the license has been correctly applied). Two Rights Statements also achieve this: No Copyright and No Copyright – United States. As indicated by the name, the latter statement only applies to the United States. By contrast, No Known Copyright signals that the legal status is inconclusive, yet the GLAM has reason to believe the work is no longer protected by IPR. In this case, the onus is on the user to ensure reuse complies with national law.
Current practices reveal confusion and misapplication of licenses versus labels by GLAMs, but also by users. This is particularly true when the media is uploaded to Wikimedia Commons, or another platform, and the rights statements conflict with the work’s actual IPR status. One area OpenGLAM might focus on is standardizing vocabularies and how rights information is recorded to Wikimedia Commons, as discussed further below.
Finally, the GLAM ecosystem needs to develop a greater awareness around appropriate digitization and collections management, including when the application of licenses, tools, or labels may be harmful or oppressive. This is discussed further in New Areas of Focus. Here, we’ll briefly note two frameworks specific to digitization and the stewardship of traditional knowledge and traditional cultural expressions to illustrate how they operate around the gaps and navigate various issues caused by our intellectual property regime.
Local Contexts is an initiative that provides labels, licenses, and notices for the traditional knowledge and intellectual property of Native, First Nations, Aboriginal, Inuit, Metis, and other Indigenous communities.27 Developed by Jane Anderson and Kim Christen, the TK Licenses and Labels aim to support Indigenous peoples’ rights to maintain, control, protect, and develop their own heritage specifically within the digital environment according to the principles established by the 2007 UN Declaration on the Rights of Indigenous Peoples.28 Similar to Creative Commons licenses and tools, the TK Licenses and Labels map onto intellectual property frameworks and enable individuals and communities to specify permissions or restrictions according to their relationships with the materials and how they should be used (or not) in digital environments, whether by those within the community or outside of it.29
In New Zealand, staff at Auckland Museum designed a cultural permissions policy specific to Māori taonga (treasures30), which sits outside of intellectual property frameworks.31 The museum’s online collections management takes an “open by default, restricted by exception” approach, while viewing images of Taonga Māori within an open adjacent framework and reuse requests subject to cultural permissions.32 This includes a set of four guidelines for managing images and their connection to the taonga both for online display and upon a request for use, depending on its purpose (e.g., commercial or noncommercial, national or non-national, a personal connection to the taonga, and so on).33 As Zoë Richardson explains, the process requires due diligence, time, constant reflection, and active collaboration with iwi (tribes) and whānau (family, both nuclear and extended) to ensure a more nuanced paradigm will guide practices irrespective of digital trends.34
At the heart of these examples lies the recognition of Indigenous peoples’ inherent sovereignty and rights to self-determination with a focus on localized approaches to digital collections and rights management. Similar initiatives for engagement with both material and digital collections are discussed in New Areas of Focus, particularly in Decolonization and Indigenization.
These details create conditions for ample confusion to arise during collections management, leading to conflicting statements and mislabeling. This can happen when we conflate the IPR status(es) of distinct layers in an image, lack information around the source work or another layer, or misunderstand what the appropriate tool, license, or label is to apply.35
Let us illustrate this with an example of a digitized archival photograph of a 19th-century vase, like the one below.36 That digital file includes (at least) three layers and rights assessments:
(1) The digital layer: the scan or photograph made by the GLAM;
(2) The analog layer: the photograph or slide in the GLAM archive; and
(3) The object layer: a 3D vase.
For a rights assessment, we would start with the source layer (3) and work our way up:
(3) The object layer: the vase was made ca. 1825, so clearly in the public domain: no IPR;
(2) The analog layer: records show the photograph was taken in 1971: copyright is likely, but the photograph is an orphan work;
(1) The digital layer: the GLAM claims no rights in the digital layer.
Despite copyright in layer (2), the GLAM’s risk assessment reveals little risk to making the image available online. It is highly unlikely a rights holder of the photograph will ever come forward. The image’s value is in its relationship to the photo archive, the vase itself, and its wider educational potential, not in its (non-existent) potential for licensing. Moreover, the GLAM sits in a jurisdiction with copyright exceptions for GLAM digitization, as well as user rights exceptions for certain personal, non-commercial, or educational uses. A decision is made to make the image available online. But how should reuse be restricted, given its non-public domain status?
After extensive internal and external consultations, the GLAM applies a CC BY-NC license and makes the digital image available online. This license is selected to protect: (1) the (potential) rights holder, whose market for commercial licensing has not been damaged if they come forward; (2) the GLAM, who may be liable to that rights holder and is not comfortable (or able) to apply an open license permitting commercial use; and (3) the user, who benefits from access to the image and the clear indication it can be used for only non-commercial purposes. In the unlikely event a request for commercial licensing should come forward, the GLAM will license the image and require the user to indemnify the GLAM against any liability via the licensing agreement. In the end, the image is now online, the public can view and reuse it for non-commercial purposes with attribution to the GLAM, and everybody wins.
As demonstrated by a Kennisland study on the accuracy of rights statements on images aggregated by Europeana, this practice is widespread.37 But that does not make it legal (don’t panic). This is because no one but the rights holder of the 1978 photograph may apply a license to its reproduction. Instead, the appropriate label to apply is either “In-copyright,” if using Rights Statements, or “@ All Rights Reserved,” (to the unknown rights holder) if not.
The desire to participate in “open access” drives the decision to apply CC BY-NC. But this desire cannot override obligations to a legal system that protects the rights holder’s interests, as well as the public’s use via the exceptions and limitations provided by national copyright law. On this point, three issues arise: first, CC BY-NC is not open access for the reasons discussed above; second, the application of any CC license (or tool) in this case violates copyright law because the GLAM holds no rights in the analog photograph;38 and, third, the copyright exceptions and limitations the GLAM seeks to align with already permit certain uses despite even a full reservation of IPR (though these benefit only national users). The effect is to undermine understanding and introduce wider uncertainty around how copyright and CC licenses actually operate, especially as technologies for reproduction advance to 3D and other new formats.39
More guidance is needed on the application of open licenses to GLAM-generated materials (i.e., CC BY and CC BY-SA). While this includes initially asking whether rights arise and therefore whether open licenses can be used, it is also important to discuss if they should be used – and when.40
As discussed in Barriers to Open Access, using CC BY and CC BY-SA allows GLAMs to maintain attribution to the host institution, an important principle in scholarship and protecting context. The desire for attribution may be legitimate. But relying on CC BY and CC BY-SA to activate this narrow obligation raises a number of issues.
First, if no IPR arises, the application of CC BY or CC BY-SA licenses will have no legal effect.
Second, if IPR does arise, GLAMs might evaluate whether these open licenses are practical. For example, with collections data and metadata, national laws may recognize copyright in various components of that data, like more descriptive text or curatorial opinions.41 Legislation might also recognize copyright in how the data itself is arranged, in addition to sui generis rights in a database,42 or related rights in composite works43 and other technologically-produced media.44 Selecting CC BY may be legally justified and reassuring to GLAMS that attribution will follow the data online – in theory.
In practice, attribution failures happen all the time. If there is no intention to ever legally enforce the copyright, why claim it? That legal obligation to attribute the data can pose significant issues when aggregated and combined with other datasets. For many, this is technically impossible. Accordingly, CC BY and CC BY-SA will chill or even prevent reuse of data. Instead, GLAMs might educate users about citation best practices in line with public missions and release data via the CC0 tool.
Third, CC BY and CC BY-SA data cannot be integrated into public platforms like Wikidata, which requires CC0. These GLAMs will lose out on exposure through one of the most visited websites in the world, and risk relevance as Wikidata advances and becomes richer and more widely used for research and networking knowledge.
Finally, because open licenses are based in copyright, restrictions on reuse will last significantly longer than the data format will be technically useful for. The legal obligation to attribute the work can therefore render the data incompatible with a range of modern applications.
A clear statement and consensus is also necessary on this point. In the EU, Article 14 aims to formally prohibit GLAMs from claiming IPR in any non-original media generated during reproduction. Yet, evidence already exists of European GLAMs claiming IPR and applying CC BY and CC BY-SA to non-original reproduction media in contradiction with national copyright law. Considering the lack of lawsuits to date, GLAMs may continue this practice and forego ever enforcing the right.45
It feels uncomfortable to forego claiming any legal right or imposing an obligation for the context to follow the public domain work online. But poor citation practices are already rampant. A claim to copyright and CC BY license will not ensure attribution to the author or institution is given. Instead, GLAMs should release media as CC0 and focus on embedding rich metadata46 and educating users around best practices for attribution and maintaining educational context.47
In fact, there are benefits to applying open tools rather than licenses to images. In their 2018 study on the societal value of images on Wikimedia Commons, Kris Erickson, Felix Rodriguez Perez, and Jesus Rodriguez Perez, found an inverse relationship to the distribution of licenses when comparing the status of uploaded images to those used downstream.48 While around 73% of the 10,000-image sample were CC BY and CC BY-SA (15.94% and 56.84%, respectively), that attribution or share-alike requirement significantly reduced the odds of the image being used externally, compared to the public domain files (which comprised only 18% of the sample).49 This suggests a preference for public domain materials, potentially due to the clear no-risk implications of downstream reuse on external platforms.
These uncertainties end up driving a wedge between initial access-driven goals and formal GLAM policies as conveyed to (and subsequently interpreted by) the public at large.
One example of this can be found on Wikimedia Commons among users who selectively respect CC BY or CC BY-SA versus remove the copyright claim and license entirely, according to the Wikimedia policy (which states in the United States, reproductions of public domain works do not receive new rights). Of course, such images may be sourced from places other than the GLAM’s digital collection to be uploaded to Wikimedia Commons, or uploaded by the GLAMs themselves.50 However, where GLAMs reserve all rights in digital surrogates, examples can be found of uploads with the images marked “public domain.” Where GLAMs release digital surrogates as CC BY or CC BY-SA, examples can be found of uploads with the Creative Commons open license intact. Things are further complicated when the work depicted is three-dimensional, rather than two-dimensional. For example, the underlying sculpture may be in the public domain, but the digital surrogate will be released CC BY-SA. The layers of rights involved may be clearly displayed in the item page details, but can become conflated when the data is linked and integrated into other platforms.51 Structured data for the Wikimedia Commons item with item level rights management, educational materials, and guidance for more accurate uploads could be improved to reflect this nuance and improve standardization.52
In fact, following 7 June 2021, the transposition deadline for Article 14 of the Copyright in the Digital Single Market Directive,53 Wikipedia’s policy for public domain artworks could be updated to reflect both the US and EU legal status for reproductions of visual artworks. OpenGLAM might develop new guidance on reuse of public domain artworks to consider the layers of digital media and how to clearly convey the IPR status of a file versus the artwork.54 Build off of ongoing work by Wikipedia communities and residents to provide support and capacity-building for GLAM staff and user-groups.
Lastly, issues also arise when GLAMs contain images of public domain works that are held in the collections of another GLAM. Let’s call the GLAM with the images “GLAM-1,” and the GLAM holding the public domain work it depicts “GLAM-2.” GLAM-1 might have the images because it made a reproduction when GLAM-2 loaned the work to GLAM-1 for an exhibition. Or, GLAM-1 may even have received a donation of photographic reproductions of many works owned by GLAM-2, and other GLAMs. Even if GLAM-1 owns the rights in those photographic reproductions or views them as public domain, GLAM-1 may be reluctant to release its photographic reproductions as open access if GLAM-2 claims copyright in its own reproductions. This is because there may be a perception that GLAM-1’s open access images will compete with GLAM-2’s in-copyright images and undermine their licensing market. Respect for GLAM-2 and maintaining the relationship will override the desire to release the public domain materials under open frameworks.
Copyright confusion and risk-averse processes already impact the management of in-copyright materials.55 The practices described in this Pub are now creating all kinds of new legacy issues for future GLAM staff and users. As digital engagement accelerates, the GLAM ecosystem needs simple and reliable methods reflecting accurate rights statuses when managing, releasing, and reusing public domain collections.
The urgency for standardized methods grows as media becomes more advanced and more rights become embedded in a given file, potentially leading to more orphan works in the future. GLAMs should be able to anticipate now how all of this will affect the IPR management of future creative media, interoperability, and ascendant technologies when data is combined and built upon to create new works. What inevitable mess awaits future curators and rights clearance? What new challenges will sharing platforms face in conveying complicated rights statements layers? If there is no real plan to actually enforce the rights claimed in public domain reproduction media, why keep claiming them?
Against this backdrop, recommendations for a Declaration include:
No new rights in non-original reproduction media;
Application of only PDM or CC0 to these materials, highlighting the benefits of CC0 for legal certainty and cross-border reuse;
Clear statements in a machine readable form embedded in data and clearly stated on GLAM websites, and support for new machine readable statements where there is demonstrated need;
Commitments to voluntarily release high-quality digital surrogates (2D and 3D), collections data, metadata, paradata, and other non-original or basic materials prepared in the context of day-to-day operations;
A definition of “open” that aligns with international definitions allowing commercial use;
Application of open licenses are appropriate only when GLAM owns rights (and in all layers of the material);
CC BY and CC BY-SA are discouraged if attribution is the only concern;
Build good citation and attribution practices into digital interfaces or websites and embed educational context in the metadata;
Expand these practices to Wikimedia Commons and Wikidata interfaces;
Forego imposing new limitations through contract-based agreements;
Incorporate licensing and labeling into practice and provide ongoing training (not one-off activities);
And what else?
These are simple propositions, but require a significant shift in mindset. They are put forward here for discussion by the wider GLAM ecosystem. But, whatever we decide, everyone must agree to sign on and apply them consistently.
An immeasurable benefit awaits upon introducing clear standards and practices to the rights management of layers that GLAMs do have control over, and the levels of ethical access extended to a global public.
Continue to Power Inequities