Statement of Best Practices in Fair Use of Collections Containing Orphan Works for Libraries, Archives, and other Memory Institutions

December 2014
Coordinating Organizations:

The Program on Information Justice and Intellectual Property, Washington College of Law, American University

Center for Media & Social ImpactSchool of Communication, American University

The Berkeley Digital Library Copyright Project, University of California, Berkeley, School of Law

Co-Facilitated by:

Patricia AufderheideCenter for Media & Social Impact, School of Communication, American University

David R. HansenUniversity of California, Berkeley, School of Law & University of North Carolina at Chapel Hill School of Law

Meredith JacobProgram on Information Justice and Intellectual Property, Washington College of Law, American University

Peter JasziProgram on Information Justice and Intellectual Property, Washington College of Law, American University

Jennifer M. UrbanSamuelson Law, Technology & Public Policy Clinic, University of California, Berkeley, School of Law

With funding from:

The Alfred P. Sloan Foundation


Libraries, archives, institutional custodians of record and other non-profit organizations that preserve memory serve as stewards for a large share of the world’s cultural, historical, and scientific record. While performing many distinctive functions and often working within larger organizations, the professionals who dedicate themselves to preserving memory also share common purposes and challenges. In this document, we refer to them collectively as “memory institution professionals.”

These professionals’ individual objectives derive from the shared mission of their institutions. Preserving a treasury of primary resources consisting primarily of unpublished documents, ephemera, and other unique items, and providing access to it for research, scholarship, and the advancement of knowledge are core features of that mission. When these resources are both securely housed and widely available, an important social interest in facilitating researchers’ and the broader public’s understanding of the knowledge contained in these collections is fulfilled. Digital technology gives memory institutions an opportunity to safely store their collections in ways that also create opportunities to give ever-greater numbers of people the benefit of them. By this means, the institutional objectives of the organizations in which memory collections are located (from larger research libraries to small specialized archives) are also advanced.1

Some collections held by memory institutions consist primarily of items old enough that copyright presents few practical challenges. Other collections may be sufficiently homogenous, and closely enough associated with a particular source, that it makes good sense to seek rights clearances before proceeding to digitize or make them accessible. A number of collections, however, lack such homogeneity, and include items from many sources. In particular, many collections include numerous “orphan” works, which are difficult or impossible to associate with active rightsholders who might give permission for their use.2 Some rightsholders may have been corporate entities that have ceased to exist. Other rightsholders may once have been locatable, but have become difficult or impossible to find today. Some likely did not create works with copyright in mind, and had no reason to remain available for inquiries.

This document considers the role that the doctrine of fair use may play in helping to resolve the copyright dilemmas that dealing with such collections can present. It addresses specifically how libraries, archives, museums, and other memory institutions can proceed with respect to collections that, based on professionals’ expertise, clearly appear to contain significant numbers of orphan works.

Why This Document Was Created

This document responds to memory institution professionals’ concerns about copyright liability in digitizing and providing digital access to collections that are believed to contain orphan works in significant numbers.

Memory institution professionals worry about orphan works because they fear the potential liability that a copyright infringement lawsuit could bring. The problem has become more significant as copyright has been extended over the last several decades while formalities of registration have been relaxed. Today, most works created in the last century are presumed copyrighted, and most copyright holders are difficult to find.

Memory institution professionals face special challenges when seeking to reuse orphan works. Archival collections, for example, often contain unpublished materials, such as historical and family photographs or notebooks, that have no markings of ownership, are not registered with the Copyright Office, and are not identified in any publicly-searchable database. The identifying information that is available can be inaccurate, incomplete, or outdated.3 This problem is compounded by the fact that, for many works commonly found in memory collections, such as scrapbooks and photographs, authors were neither motivated by nor even aware of the copyright protection that automatically attached to their creations. The likelihood of those authors maintaining contact with their works is small. Further complicating the situation are limitations on institutional resources that make comprehensive, item-by-item clearance requests prohibitive, even when theoretically feasible.

Memory institution professionals commonly manage collections containing materials that are, practically speaking, impossible to identify and seek copyright permission for, item by item. If they fail to address copyright clearance issues, they could compromise their institutions’ public missions. Nevertheless, faithful representation of a collection in its entirety could be critical to fulfilling an institution’s missions to preserve the past and to make research materials available, including online. Even in those situations where mission may be served by active selection and arrangement of materials, avoidance of copyright issues should not be the primary consideration influencing institutional choices about inclusion and presentation.

In the United States in the last decade, the orphan works problem has generated proposed legislation and several Copyright Office studies.4 But none of those bills passed and there has been no further legislative relief or guidance.

Over the same period, individual archivists and professional organizations, such as the Society of American Archivists,5 have documented ways to locate rightsholders of copyrighted materials, focusing on capturing a complete set of options for investigating the legal status of individual works. However, until now there have been no clear statements about what librarians and archivists consider professionally and practically appropriate practices regarding orphan works embedded in memory institution collections. Neither has there been any documentation of this community’s views on how fair use applies to the management of such collections.

How This Document Was Created

To address this need, the co-facilitators of this document undertook a multi-stage process to help the community discover and express its collective views on the issue of orphan works and how to address them.

In September 2012, the co-facilitators convened a workshop in Washington, D.C. to discuss the orphan works problem and how to address it. The workshop brought together more than 30 librarians, archivists, and other memory institution professionals from a wide range of organizations. Both large and small institutions participated, including academic and governmental libraries and archives, public libraries, private museums and archives, local historical societies, web archiving projects, and the Digital Public Library of America, along with representatives of organizations that support the work of those institutions, such as the Association of Research Libraries and the American Library Association. Participants came from a range of geographic locations throughout the United States.

The views of the workshop group were combined with the results of several in-depth follow-up interviews and documented in the Report on Orphan Works: Challenges for Libraries, Archives, and Other Memory Institutions.6 A primary need identified in the Report was to develop

[B]est practices to help guide and empower digitizing institutions that seek to make good faith efforts in using orphan works. This should include best practices for topics including identifying when a search is desirable, the form that a search should take in various circumstances, the role of ancillary considerations (including privacy) in designing any search, and approaches to seeking permissions for use when a search has indicated a possible copyright owner.7

With that goal in mind, project members undertook a series of intensive deliberative group meetings, holding ten half-day sessions in nine U.S. cities.8 The discussion groups were primarily composed of special collections librarians and archivists. The groups included 72 memory institution professionals whose primary job was to acquire, maintain, and provide access to archival and special collection materials. An additional 20 served a similar role, but with a focus almost entirely on the digital environment. The groups also included 10 memory institution professionals who served in management roles where they encounter copyright and orphan works challenges, 10 who have special experience providing copyright assessment and advice for library and archives collections, and 17 who serve a variety of roles, often within smaller organizations. A large number of these participants—79 in total—worked with collections at university libraries, though the size, resources, and type of collections held by those institutions varied widely. Others included 9 participants from state and federal libraries and archives, 11 participants from museums, 19 participants from independent (typically small) historical societies, archives, and libraries, and 11 participants from public libraries.

While these discussions began with a focus on orphan works as such, deliberative groups quickly made clear that the nature of most digitization projects—especially those on a larger scale—meant that librarians and archivists had little practical way of addressing individual orphan works separately from the collections of which they are a part. As a result, the conversation in most deliberative group sessions soon shifted to how fair use supports the digitization of collections that are believed to contain orphan works.

Thus, while the resulting statement offers guidance to memory institution professionals in making decisions about orphan works, it also considers mass digitization projects more generally—just as did the participants in our meetings.

Overall, the deliberative group conversations revealed a deep understanding of how institutional mission, the nature of the collection, and the potential for harm to unidentified stakeholders all affect a fair use determination. The groups’ consensus-based, middle-ground understanding of how the fair use doctrine applies is documented below in these best practices.

To ensure that the application of fair use reflected in this statement falls within the bounds of reason and precedent, an independent legal advisory panel of recognized copyright experts has reviewed this document. This should not, however, be construed as representing their legal advice. Rather, the document represents the voice of the community of memory institution professionals itself.

What This Is

This statement of best practices represents the carefully derived community consensus view that emerged from more than a year of discussions. It articulates two broad best practices principles, and, accompanying the second, a range of related qualifications that describe practices supported by fair use that promote the mission of memory institution professionals who manage collections that are believed to contain orphan works. The principles should not be read without the associated qualifying language.

The best practices are stated in general terms so that any institution can take these best practices and apply them to its own circumstances, as a tool to help inform its own legal and risk analyses.

The scope of this document:
Although the principles and best practices articulated in this document might logically be extended to collections that contain copyrighted material more generally, this document, like the discussions that led up to it, is specifically intended to guide activities with collections that are believed to contain a significant number of orphan works. Orphan works have several distinct characteristics that can be relevant to the fair use analysis, as discussed below. Most importantly, they are, by definition, not active in the market and may not have been prepared for market purposes in the first place. This increases the social value of digitizing them while simultaneously limiting any resulting economic harm to rightsholders.

What This Isn't

This statement does not describe the limits of fair use in memory institutions, but articulates how professionals understand the doctrine to apply in a limited number of recurrent situations that they face in digitizing collections that include orphan works. Institutions may be able to make persuasive arguments for fair use in circumstances that go beyond the shared
norms expressed here, just as they may have good reasons for choosing policies that do not take full advantage of these consensus principles.

This statement was not negotiated with rightsholders that do not have as their mission to collect, preserve, and provide access to collections of material.9

This statement lays out the reasoning by which memory institution professionals believe they can exercise their fair use rights; it does not present rules or bright-line tests. Rules and bright-line tests ignore the fact that, as is true of all rights that enable freedom of expression, the effective exercise of fair use requires both an understanding of the particular context and consistent reasoning from instance to instance.

Nor does this statement suggest that the exercise of fair use rights is an obligation of memory institutions. Relationships with donors, extraordinary privacy concerns related to some collections, and each institution’s own tolerance for risk are just some of the factors that may lead some to decide not to exercise fair use rights in certain situations.

This statement is not a guide to using material that people give the public permission to use, such as works covered by Creative Commons licenses. While fair use applies to such works, anyone may use those works in ways their owners authorize, in addition to ways permitted by the fair use doctrine.

Similarly, it is not a guide to the use of works that are in the public domain; copyright does not limit the use of these works, including uses that otherwise would far exceed the bounds of fair use. Because many library and archives digitization projects may include early 20th-century works that are potentially in the public domain, we repeat a recommendation made in the Report on Orphan Works Challenges10 for the community to develop educational materials to help librarians and archivists make public domain status determinations.11

Risk Management and Fair Use Analysis:
Digitization projects in memory institutions naturally give rise to risk management questions. Indeed, throughout the discussion groups, memory institution professionals frequently expressed ideas about risk management strategies that might be appropriate in using collections that are believed to contain orphan works. Managing institutional risk in the use of copyrighted collection materials is important, and archivists and librarians have recently produced several helpful guidance documents on this subject.12 Those documents emphasize that the fair use assessment is an important part of any risk management strategy.

This document, however, is more narrowly focused. It reflects memory institution professionals’ consensus discussions of situations where institutional practices support a claim of fair use. Accordingly, it emphasizes clearly articulating how the law of fair use can apply to memory institutions’ use of collections that are believed to contain orphan works. Fair use, of course, is a “rule of reason.” This means that some of the factors that directly affect the fair use assessment may also affect the level of risk experienced by an institution for other practical reasons. In these cases, practices may positively affect the user’s position under fair use, while at the same time decreasing the likelihood that any rightsholder will come forward to challenge the digitization project.

Fair Use Backgrounder

The goal of U.S. copyright law is to promote the progress of science, the creation of culture and art, and the dissemination of ideas. Its best-known feature is protection of owners’ rights. But copying, quoting, and generally re-contextualizing and re-using existing cultural material can be, under some circumstances, a critically important part of facilitating research, promoting new discoveries, and enabling the creation of new creative works. In fact, the value of these practices is so well established that it is written into the social bargain at the heart of copyright law. Our society offers some exclusive
rights in copyrighted works to their creators to encourage them to produce science and culture. At the same time, we limit those rights in several important ways so that the primary intended beneficiary of copyright law—the public—can benefit from those works.13 One significant limitation is that the rights expire; copyright lasts for a limited time, and then works enter the public domain, where they are free for use by all. In practice, of course, recent extensions of copyright term have blunted the significance of this one potent limitation. Likewise, Section 108 of the Copyright Act allows for certain preservation activities by libraries and archives, but fails to address many issues that memory institution professionals face on a daily basis.14 Other limitations, including fair use, allow more generally for the use of works still protected by copyright without permission or payment. Without those uses, the public could lose out on important new works and both public discourse and public memory could be impoverished.

Reflecting this societal value, copyright law is particularly solicitous of library and archives uses. For example, Section 504(c)(2) of the Copyright Act shields non-profit libraries and archives from costly statutory damage awards when they make a good faith assertion of fair use, even if ultimately an incorrect one. Additionally, many of the uses libraries and archives make not only fall in line with fair use principles, but are also singled out by Congress in specific sections of the Copyright Act. These include Section 108, which identifies particular circumstances in which libraries can always make copies for preservation and for activities such as interlibrary loan; Section 110, which permits specific types of non-profit educational public performances; and Section 121, which allows libraries to make accessible copies for patrons who have print disabilities. Fair use complements and supplements these sometimes narrowly-drawn specific limitations.15

Fair use is the most important and well-established of the copyright limits; it has been part of American copyright law for more than 170 years. Congress has recognized fair use as a doctrine with specific application to libraries and archives. Section 107 of the Copyright Act, which codified the fair use doctrine in 1976, specifically refers in its preamble to activities supported by libraries and archives, such as “criticism, comment. . . , teaching . . . , scholarship, [and] research.” The legislative history goes further and explains, by way of example, that in the context of older films, “efforts . . . to rescue and preserve this irreplaceable contribution to our culture are to be applauded, and the making of duplicate copies for purposes of archival preservation certainly falls within the scope of ‘fair use.’”16 The statute directs judges to take account of several considerations in analyzing fair use: the purpose and character of the use, the nature of the work used, the extent of the use, and its economic effect (the so-called “four factors”). In addition to this non-exclusive list, some judges take into explicit account the ways in which a challenged use may promote public access to information. Others, as will be explained, incorporate this consideration into their analysis of factors themselves.

Because copyright law describes the right of fair use in general terms, the fair use doctrine can adjust to evolving circumstances, thus continuing to fulfill the role of promoting culture. Rather than following a prescriptive formula, judges and lawyers assess whether a particular use of copyrighted material is “fair” according to an “equitable rule of reason.” This means taking into account all facts and circumstances to decide if an unlicensed use of copyrighted material generates social or cultural benefits greater than the cost imposed on the copyright owner.

Judicial decisions regarding fair use give the strongest guidance about how to apply the doctrine. In analyzing the factors, the history of fair use litigation since the early 1990s shows that judges return to two key analytical questions:

1. Did the use “transform” the material taken from the copyrighted
work by using it for a broadly beneficial purpose different
from that of the original, or did it just repeat the work for the
same intent and value as the original, in effect substituting for it?

2. Was the material taken appropriate in kind and amount, considering
the nature of the copyrighted work and of the use?

These two questions effectively collapse the “four factors.” The first addresses the first two factors (nature of the use and nature of the work used), and the second rephrases the third factor. Both key questions touch on the fourth factor, whether the use will cause excessive economic harm to the owner. If the answer to both questions is “yes,” a court is likely to find a use fair, even if the work is used in its entirety. Because that is true, the risk of a challenge is dramatically reduced.

These questions help weigh the balance (benefit to society versus cost to rightsholder) at the heart of the fair use analysis. Fair use ensures that rightsholders do not have a monopoly over transformative uses of their works. It also protects rightsholders from infringement. When a use merely supplants a copyright owner’s core market rather than having a transformative purpose, it is unlikely to be fair. Thus, for example, a library cannot acquire currently marketed materials for its collection simply by photocopying or scanning published editions.

Cases brought in recent years have generally favored library and archives uses, though several of those decisions are currently being appealed. Few fair use cases specifically involve libraries, archives, or other memory institutions. In general, however, case law since the early 1990s generally favors conscientious fair users. They have held that for a use to be considered “transformative,” it need not be one that modifies or literally revises copyrighted material: uses that repurpose or recontextualize copyrighted content in order to present it to a new audience for a new purpose can qualify as well. Courts have found that full-text copying facilitates information discovery.

For instance, Google’s image search, and Google Books and HathiTrust Digital Library’s book search functionality are transformative and fair uses.17 Court decisions also show that the more coherent an account the user can give of how and why the material was borrowed, the more likely the use is to be considered transformative. The flexibility of fair use can sometimes lead users to wonder if particular contemplated uses fall within the scope of the doctrine, and even to wish for clearer rules or brighter lines. But fair use’s flexibility is also its strength. Legal scholars who are experts in fair use have concluded that the doctrine is both predictable and consistent, especially when applied in certain recurrent situations.18 Even without case law specifically addressing a use, judges and lawyers consider expectations and practice—whether the user acted reasonably and in good faith in light of standards of accepted practice in his or her particular field.19 Therefore one way of creating a better understanding of what fair use permits is to document the considered attitudes and best practices of a professional community as it works to apply fair use.

Nine other communities of practice have developed best practices in fair use for themselves since 2005,20 and all have benefited from establishing a community understanding of how to employ their fair use rights, with little or no resulting controversy. Thus, for example, teachers employ fair use to create multimedia lessons for their students, while journalists rely on it to support their use of excerpts and quotes in their news reports. Filmmakers use, reinterpret, and critique copyrighted materials in their films. Academic and research librarians rely on fair use to support a wide variety of teaching, preservation, and general public access uses. Indeed, the existing code of best practices that overlaps most with the concerns of memory institution professionals is the Code of Best Practices in Fair Use for Academic and Research Libraries.21

Where a memory institution elects to rely on fair use in support of an aspect of its mission, this right is exercised most usefully by behaving consistently throughout the lifecycle of related professional activities. Reliance on the right of fair use is most compelling when reasonable steps to avoid rights conflicts and concerns have been employed from the outset. Thus, for example, including in donors’ agreements a transfer of or license for whatever intellectual property rights they may hold along with any gift of physical materials makes many downstream decisions much easier. So does specifying digitization in a donor’s agreement.22 Likewise, cataloguing and collection storage that exclude some sensitive materials and consider privacy, as discussed below, can, given the resources of the organization and the size and nature of the collection, facilitate more confident employment of fair use.

Finally, memory institutions that employ fair use responsibly to make material available, physically or virtually, have no general duty to police, monitor, or regulate the activities of those who make use of the resource. With this in mind, memory institution professionals consulted for this document nearly unanimously rejected the notion that they should use digital rights management technologies routinely in their practice.

Fair Use Applied

Courts have recognized research access to primary sources as an important value to be furthered by the copyright fair use doctrine.23 In particular, three characteristics of library and archives uses described in the principles and best practices below support fair use in collecting, preserving, and giving access to collections containing orphan works:24

1. The uses described in these practices minimize any potential economic harm to owners of copyrighted works, presenting a strong case under the fourth fair use factor, the impact on the market for the work. In particular, the use of orphan works by themselves poses almost no risk of market harm; if owners cannot be located, there is little chance that there is a current, functioning market for those works. Further, the fact that those works remain unexploited is a strong signal that their potential economic value in any future-developed market would be small.25

More generally, memory institution professionals expressed a shared concern for minimizing the chance of economic harm to rightsholders with several of the best practices set forth below. This includes seeking copyright clearances when reasonable and practicable, such as when significant clusters of works are readily traceable to a single rightsholder, when contacting rightsholders can be automated, when works have readily identifiable and significant market value (e.g., works associated with a high-profile individual), and when works in the collection are predominately new—created within the last 25 years. The best practices also favor selectively excluding certain works when appropriate, and limiting access in certain circumstances. Similarly, these best practices favor minimizing market harm by asking memory institution professionals to incorporate what they learn about their collection through acquisition, cataloging, other curation, and dialogue with the public, into their decisions to make works available and decisions to continue making works available.

In sum, these best practices document a methodology by which implementing organizations can enhance their fair use position by weeding out exceptional items as to which their fair use case is weaker because it might present more risk of market harm.

2. These uses are generally transformative, repurposing, or recontextualizing material in a way that has an overall societal benefit. Consistent with their mission to preserve and provide meaningful access to their collections through digitizing and displaying digital archives, archivists and librarians value both curated collections (for example, a collections of materials, drawn from a larger universe of material, selected and arranged to tell a story) and undifferentiated collections (for example, collections that—to preserve a record and accurately portray historical events—are presented in their entirety and in their original arrangement).26

Some memory institution professionals, for example, those working with special collections that include materials related to a significant historical event, explained that they routinely select, arrange, and contextualize material together in a way that transforms material because it generally uses it for a purpose and to convey a different meaning than the original work or works by themselves. Repeatedly, transformativeness arguments were articulated for each approach. Some memory institution professionals, for example, those working with special collections that include materials related to a significant historical event, explained that they routinely select, arrange, and contextualize material together in a way that transforms material because it generally uses it for a purpose and to convey a different meaning than the original work or works by themselves. But even online collections that include all available material and are presented without altering the arrangement of the original record keeper—as is common practice in archives—are more than the sum of their parts; presenting the collection together, as well as its further curation, are acts that serve a new, transformative purpose.

Orphan works are embedded in the fabric of broader collections; filtering them out would not only be impractical, but would also impair mission in general and interfere with effective curation. Indeed, memory institution professionals considered the digitization of archives and special collections for preservation and access to be a fair use activity, when done appropriately, with respect to all items included, not just orphan works.

3. The uses are made in good faith by libraries and archives in the course of performing their mission in the public interest. Several of the best practices below aim to promote responsible uses that are respectful of other stakeholders’ interests—including, for example, privacy concerns.27 Although the statutory factors predominate in the fair use jurisprudence, case law certainly supports the idea that consideration of “good faith” can appropriately enter into fair use analysis.28 The strong consensus of memory institution professionals who contributed to this statement is that the most comfortable position from which to assert their fair use rights was one in which they could defend their activities both legally and ethically.

Institutional capacity to adopt the best practices described below will vary widely. The professionals whose views are reflected here did not believe that the adoption of any (let alone all) of the practices itemized below was a precondition for the exercise of fair use by memory institutions. They did conclude that making reasonable efforts to adopt such practices, considering each institution’s circumstances and resources, will help them assert fair use rights most successfully.

Memory institution professionals described two overarching and recurrent copyright issues in their practice:

      • Preservation: Whether or not to digitize collections for preservation purposes when collections contained significant numbers of items, the copyright ownership of which could not reasonably be determined. They felt a high degree of confidence that fair use did and should apply to such preservation activities without significant qualification.
      • Access: Whether such digitized collections could be made available for research and study, both within the physical institutions, and also via online access. They strongly shared the conclusion that fair use generally applied in this situation, but with nuances in the consensus reflected below.

Code of Best Practices

ONE: Preservation


Memory institution professionals strongly believed that preservation was foundational to any other mission they could perform,
and that fair use was available to them for this purpose, in addition to the specific preservation exceptions of 17 U.S.C. Section 108.29


Fair use supports the digital preservation of materials in archival and special collections, without regard to their status as orphan works.

TWO: Access


Memory institution professionals believed that providing access to their collections was the ultimate goal of all institutional activities, including preservation. They believed that fair use was available to them for this purpose, but they also recognized both the idiosyncrasies of particular collections, and a common set of issues that a responsible professional must consider before allowing access. They believed that certain best practices for acquisition, cataloging, and management were important to the responsible exercise of their fair use right.

Fair use is a fact-specific doctrine, and memory institution professionals believed that these additional best practices should inform professional decision-making about what measures were appropriate in any particular circumstance. Not every practice will be relevant for every contemplated access project, in light of the institution’s mission and other institutionally unique characteristics. The practices are presented here in chronological order of typical decision-making, not in order of importance.


Fair use supports professionals’ efforts to provide on-premises and online public access to archival and special collections
that can reasonably be expected to contain significant numbers of orphan works, including collections that include other copyrighted materials. This general principle should be applied along with the seven additional best practices described on the next pages of this document. These additional best practices relate to:

1. Acquisition
2. Clearances
3. Selective Exclusions from Access
4. Curation
5. Conditions on Availability
6. Dialogue with the Public
7. Providing Copies to Members of the Public

Additional Best Practices for Providing Access



The case for fair use in memory institutions is stronger if both procedures for the acquisition and processing of new material are structured with appropriate diligence in mind. Opportunities to make prudent decisions that facilitate fair use occur at the level of donor agreements, cataloguing and description, and storage.


Donor agreements: In cases where donors are believed to own rights, archives and special collections should, where reasonably possible, attempt to secure affirmative permissions from donors to make collection materials available, including by digital means.

Cataloguing and describing rights information:30 Wherever information about provenance and authorship is available, it should be firmly and permanently associated with the relevant items. This should happen at the folder or box level, not that of the individual item (memory institution professionals no longer expect to engage routinely in item-level processing). Memory institution professionals should:

i. Make a reasoned assessment of a collection’s overall “intellectual property profile,” including an estimate of the prevalence of orphan works, based on a sample of the whole.

ii. Adapt their systems for collection assessment to respond to new information technologies and new demands on their resources.

iii. Collect as much data as is reasonably possible about the creators and sources of the different categories of items contained in a collection.

Storage: Data security should be used for these materials, whether they are locally or remotely stored, in accordance with current general practices employed in digital collections.



Memory institution professionals strongly believed that seeking permissions to provide access to materials was desirable in some particular cases. As a matter of legal doctrine, a fair user never is required to obtain a rightsholder’s prior consent, but under some circumstances seeking such consent may be advisable to document the absence of market harm. Indeed, in order to sustain the strength of memory institutions’ fair use claims generally, many professionals believed it was desirable in certain special cases to seek permission before providing digital access. They believed that in particular instances they should refer to the best professional judgment of the community, reflected in the best practices below, to determine whether clearances should be sought. In the event that the institution decides to seek permissions, they can wisely use the Society of American Archivists’ 2009 “Orphan Works: Statement of Best Practices” as a guide to ways and means of doing so.


Seeking permissions for use: Make attempts to secure copyright clearance in certain situations where this is reasonable, especially those characterized by the significant presence in the collection of the following
types of works:

i. Significant clusters of items traceable to a known or easily identifiable copyright owner (or groups of related owners);31

ii. Significant clusters where contacting rights owners can be automated;

iii. Individual items representing works that have readily identifiable and significant market value, including material
related to high-profile individuals;

iv. A predominance in the collection of materials created within 25 years.

Avoiding repetitive permissions requests: When requesting permissions, phrase requests broadly so as to avoid the costs and complications of seeking additional permissions for future uses.

Respecting rightsholders’ preferences: Honor the wishes of copyright holders who, once contacted, withhold permission.

Documentation: Document all actions and processes relating to permissions.



In some cases, memory institutions have collections in which there is information that offers significant challenges to public access. The professionals in deliberative groups believed that they could bolster their fair use arguments by paying careful attention to areas of particular concern, especially in two areas: privacy and third-party media. They also believed they should fully disclose online the nature of any exclusions they made, to document the fullness of the collection.

Opinions differed on the severity of the measures to be taken. Where privacy was concerned, some believed that where redaction or exclusion could not take place, professionals can consider delaying online access. Some believed that unpublished copyrighted works by high-profile literary authors, composers, and other culturally resonant figures should be held offline. Others cautioned that access to the historical record should not be impeded unduly.


Personal privacy: Balance the general interest in access to information with privacy considerations, and consider selectively excluding from broader online access, at least for a period of time, material that cannot be cleared if there are special ethical concerns, such as the privacy of people who are or are likely to be living, risk management considerations,
or both. Memory institution professionals believed they should consider the feasibility of identifying and limiting the online availability of items that may prove especially embarrassing to, or serve as significant evidence of wrongdoing on the part of, living individuals.

In all cases, reasonable efforts should be undertaken to redact appropriate information, especially where this can be accomplished by automated or semi-automated means (as, for example, with social security numbers).

Third-party media: When collections, for example personal papers, include published third-party books, journals, DVDs, media files, or other material, where possible such materials should be held offline, unless they contain significant annotations or are otherwise important to the story the collection has to tell.



Sometimes collections include copyrighted material that is of special interest or concern—for instance, a well-known figure’s correspondence. Memory institution professionals believed that in these circumstances it would be possible to employ fair use when the institution is adding value in some way. The most common method discussed was the addition of curatorial value, which can take a variety of forms. Depending on resource considerations and the nature of the collection and institutional mission, and by way of example only, a collection could do one or more of the following: limit the resource to items chosen for special significance; provide context through curatorial narratives; provide users with specialized search tools; or embed metadata with specific rights-related information.


The rationale for including potentially copyrighted material in an online resource is stronger when the institution adds significant value to the digital collection.



Memory institution professionals believed that establishing a shared goodfaith understanding with their users about appropriate use of a collection was important. They also placed a high value on users’ ability to access materials anonymously. They believed click-through contracts that prescribe appropriate use of materials, even if not enforceable, would be valuable because they attract user attention and reinforce an educational message.


First-time users of a digital resource should read and agree to basic terms of use, including promising not to use materials from the site in violation of copyright or other laws. Memory institution professionals should preserve use anonymity wherever possible.



Memory institution professionals believed that communication about collections was important to successful employment of fair use, because it demonstrated good faith and offered potential rightsholders the ability to request changes in the online presentation. Communication with the public offers ancillary benefits as well. Dialogue with users was important to them, not only to honor the value of transparency and good practices by users, but also because users can be valuable contributors and even help in providing additional information about collection materials.


Memory institution professionals should be as transparent as possible about their practices, explain why they are making a collection public, and make public background information about the project’s goals and procedures. They should also offer the public a way to give feedback, such as a prominent link on the website of the resource and a dedicated email address for the submission of questions, comments, concerns, and other additional information. Institutions should not simply promise takedowns in the event of receiving a complaint, but should always engage in dialogue as a first step. They also should provide prompt feedback to user comment and inquiry.



Many users have legitimate scholarly and other interests that are facilitated by the creation of their own copies. Memory institution professionals did not believe that providing such copies would impair their reliance on fair use. But they did believe that they should engage with users to make them fully aware of their own obligations with respect to the materials, and to demonstrate the institution’s good faith in providing those materials. End-user agreements, currently used in many physical settings, may not be effective to limit institutional liability, but memory institution professionals thought they could
be useful to demonstrate good faith toward known and unknown copyright owners. They also thought that institutions should limit access fees to recovery of costs when they facilitate user access to actual or virtual copies.


Users who seek digital copies should identify themselves and be informed of the importance of using such copies only in copyright-compliant ways, by means of an agreement or otherwise.

End Notes


1     See Michael J. Madison, Knowledge Curation, 86 Notre Dame L. Rev. 1957 (2011), available at abstract=1848086 (discussing the variety of ways that the law facilitates (and fails to facilitate) memory institutions’ efforts to fulfill this mission in the face of new technology).

2     Because the goal of this project was to articulate how fair use applies in a wide variety of contexts with respect to collections containing orphan works, it was neither necessary nor useful to rigidly adhere to a particular definition of “orphan works,” though many definitions exist. Rather, we asked deliberative group participants to use a broad conception of the term that would allow them to consider the variety of factors that bear on the decision to make use of a work in a given scenario. For a discussion of the alternative definitions, see David R. Hansen, Orphan Works: Definitional Issues (Berkeley Digital Library Copyright Project, White Paper No. 1, 2011),

3     See Maggie Dickson, Due Diligence, Futile Effort: Copyright and the Digitization of the Thomas E. Watson Papers, 73 Am. Archivist 626 (2010),; Dharma Akmon, Only with Your Permission: How Rights Holders Respond (or Don’t Respond) to Requests to Display Archival Materials Online, 10 Archival Sci. 45 (2010), (both studies reporting large numbers of works in sample collections for which rightsholders could not be located because of missing or difficult-to-obtain contact information).

4     A more thorough review of the orphan works problem can be found in the report that preceded the development of this statement, Report on Orphan Works Challenges for Libraries, Archives, and Other Memory Organizations (January 2013),

5    Society of American Archivists, Orphan Works: Statement of Best Practices (2009), standards/OWBP-V4.pdf.

6     Report on Orphan Works Challenges for Libraries, Archives, and Other Memory Organizations (January 2013),

7     Id. at 13-14.

8     These half-day sessions were intentionally diverse both in type of participant and location. Participants included small and large institutions, custodians of special and more general collections, librarians, archivists, rights management professionals, and curators. Three deliberative group sessions were held on the West Coast, two in the Midwest, three in the Southeast and Mid-Atlantic, and two in the Northeast.

9     See Kenneth Crews, The Law of Fair Use and the Illusion of Fair-Use Guidelines, 62 Ohio St. L.J. 602 (2001), available at (reviewing the history of fair use guidelines in education, and noting the difficulties facing users of restrictive guidelines and shared understanding agreements that have been created jointly with rights holders).

10    Report on Orphan Works Challenges, supra note 4, at 13-14.

11    For more, see Peter Hirtle, Copyright Term and the Public Domain in the United States (Jan. 1, 2014), available at See also Menesha A. Mannapperuma, Brianna L. Schofield, Andrea K. Yankovsky, Lila Bailey, & Jennifer M. Urban, Is it in the Public Domain? A Handbook for Evaluating the Copyright Status of a Work Created in the United States Between January 1, 1923 and December 31, 1977 (May 2014), available at

12    See Victoria Stobo with Ronan Deazley and Ian G. Anderson , Copyright & Risk: Scoping the Wellcome Digital Library Project (CREATe Working Paper No. 10, 2013),; Peter Hirtle, Emily Hudson & Andrew T. Kenyon, Copyright and Cultural Institutions: Guidelines for Digitization for U.S. Libraries, Archives & Museums 193-212 (2009),; OCLC, Well-intentioned practice for putting digitized collections of unpublished materials online (2010),; Kevin L. Smith, Copyright Risk Management: Principles and Strategies for Large-Scale Digitization Projects in Special Collections, Research Library Issues, June, 2012,

13    In Sony Corp. of Am. v. Universal City Studios., 464 U.S. 417, 429 (1984) (“The monopoly privileges that Congress may authorize are neither unlimited nor primarily designed to provide a special private benefit. Rather, the limited grant is a means by which an important public purpose may be achieved. It is intended to motivate the creative activity of authors and inventors by the provision of a special reward and to allow the public access to the products of their genius after the limited period of exclusive control has expired.”).

14    See Jonathan Band, The Impact of Substantial Compliance with Copyright Exceptions on Fair Use, 59 J. Copyright Soc’y 453 (2012), available at

15    17 U.S.C. § 108(f)(4) (2012) (“Nothing in this section in any way affects the rights of fair use and provided by section 107. . . .” See Authors Guild Inc. v. HathiTrust, 755 F.3d 87, 94 n.4 (2d Cir. 2014). (Section 108 does not foreclose analysis of defendant libraries activities on the alternative basis of fair use).

16    See H.R. Rep. No. 94–1476, at 73 (1976), 1976 U.S.C.C.A.N. 5659, 5687.

17    See Perfect 10, Inc. v., Inc., 508 F. 3d 1146 (9th Cir. 2007) (Google image search); Authors Guild, Inc. v. HathiTrust, 755 F.3d 87, 96-97 (2d Cir. 2014); Authors Guild v. Google Inc., 954 F. Supp. 2d 282 (S.D.N.Y. 2013).

18    See Pamela Samuelson, Unbundling Fair Uses, 77 Fordham L. Rev. 2537 (2009), (discussing several “policy-relevant clusters” in which fair use is consistently applied); Barton Beebe, An Empirical Study of United States Copyright Fair Use Opinions, 1978-2005, 156 U. Pa. L. Rev. 549, 574-5 (2008),; Matthew Sag, Predicting Fair Use, 73 Ohio St. L.J. 47 (2012),; Neil Netanel, Making Sense of Fair Use, 15 Lewis & Clark L. Rev. 716 (2011),

19    See Michael J. Madison, A Pattern-Oriented Approach to Fair Use, 45 Wm. & Mary L. Rev. 1525 (2004), (reviewing fair use case law illustrating how courts often implicitly, and sometimes explicitly, assess fair use questions based on whether the use falls within or beyond recognized social or
cultural patterns).

20    See Best Practices, Center for Media & Social Impact, American University, http://www.

21    Assoc. of Research Lib., Code of Best Practices in Fair Use for Academic and Research Libraries (2012),

22    Model acquisition policies are available from the Association of Research Libraries, the Society of American Archivists (SAA), and others. In particular, institutions should consider the use of rights “quitclaims,” so that donors can grant whatever rights they have without claiming that they actually have any.

23    See Sundeman v. Seajay Soc’y, Inc., 142 F.3d 194, 204 (4th Cir. 1998) (“[All defendants”] purposes serve the public benefit and aid in the development of the arts.”)

24    Jennifer Urban, in How Fair Use Can Help Solve the Orphan Works Problem, 27 Berkeley Tech. L.J. 1379 (2012),, offers a more complete and persuasive account of how these and other characteristics can support an assertion of fair use of orphan works by libraries and archives.

25    These characteristics of orphan works also make for a strong position under the second fair use factor, the nature of the work. See, e.g., Hofheinz v. A&E Television Networks, 146 F. Supp. 2d 442, 447 (S.D.N.Y. 2001) (explaining that, under the second factor, “[i]f the work is out of circulation and unavailable for purchase through normal channels, the user may have more justification for reproducing it than in the ordinary case.”).

26    In both approaches, the message is “this is what was said” rather than “this is what you should know or believe.” See The Swatch Group Management Ltd. v. Bloomberg 756 F.3d 73, 81, at *8 (2d Cir. 2014); Fox News Network, LLC v. TVEyes, Inc., 2014 WL 4444043, at *20 (S.D.N.Y. Sept. 9, 2014).

27    Privacy is a factor that courts have sometimes considered, but it has not emerged as a critical part of the fair use analysis. See Judge Jon O. Newman, Copyright Law and the Protection of Privacy, 12 Colum.-VLAJ.L. & Arts 459 (1989). Documenting provenance and authorship is another example of a best practice that demonstrates good faith. Further, efforts to limit inappropriate uses by end users, e.g., data security measures or click-on licenses, both minimize the economic harm to copyright owners and manifest good faith.

28    See Harper & Row, Publishers, Inc. v. Nation Enterprises, 471 U.S. 539, 562-63 (1985) “Fair use presupposes good faith and fair dealing.”) (citations omitted); NXIVM Corp. v. Ross Institute, 364 F.3d 471, 479 n.2 (2d Cir. 2004) (explaining why the good- or bad-faith analysis should not be dispositive to the fair use inquiry).

29    Judge Chin’s opinion in Authors Guild v. Google, Inc., strongly suggests that systematic digitization for preservation purposes is a fair use. 954 F.Supp.2d at 293. The recent Second Circuit decision in Authors Guild, Inc. v. HathiTrust implicitly reached the same conclusion. See Jonathan Band, What Does The HathiTrust Decision Mean For Libraries, The Court of Appeals did not reach the question of whether HathiTrust could provide access to digitally preserved copies for replacement purposes because the plaintiffs did not have standing to raise it; previously, however, Judge Baer’s district court opinion had identified preservation as a potential fair use activity.

30    In this conclusion and in several others regarding clearances, extra curation, and dialogue with the public, professionals echoed the “Well-Intentioned Practices for Putting Digitized Collections of Unpublished Materials Online” issued by OCLC Research. In some instances discussion group participants referenced that document, and other times they appeared to arrive at these conclusions independently.

31    An example cited by several participants in discussion groups was a body of electronic correspondence to the individual who was the focus of a collection, perhaps preserved—along with the correspondents’ e-mail addresses—on the hard drive or his of her computer.


Endorsed by:

Archives, Museum of Fine Arts, Boston –

Association for Recorded Sound Collections –

Association of Research Libraries –

Association of Southeastern Research Libraries –

California Digital Library –

California State University, Fresno, Henry Madden Library –

California State University, Northridge, Oviatt Library-

Digital Commonwealth –

Digital Public Library of America –

Emory University Library & Information Technology Services –

Empire State Digital Network –

Metropolitan New York Library Council –

NCSU Libraries –

North Carolina Digital Heritage Center –

University of California – Berkeley Library –

University of California – Davis Library –

University of California – Irvine Libraries –

University of California – Los Angeles Libraries –

University of California – Merced Library –

University of California – Riverside Library –

University of California – San Francisco Library and Center for Knowledge Management –

University of California – Santa Barbara Library –

University of California – Santa Cruz Library –

University of Florida, George A Smathers Library –

University of Georgia Libraries –

University of Illinois Urbana-Champaign Library –

University of Massachusetts Amherst, University Libraries –

University of North Carolina at Chapel Hill Libraries –

Wisconsin Historical Society –