June 2012 Research Library Issues: A QuarterlLy Report from ARL, CNI, and SPARC RLI 279 19 Copyright Risk Management Four Strategies for Evaluating Risk The first strategy for evaluating the risk associated with a digitization project is to recognize that, in many collections, at least some of the material will be in the public domain. This recognition helps us reduce, in our perception of a project, the number of risky items involved. As has been noted, the public domain is usually considered, in digitization planning, only when an entire collection is likely to be in the public domain. That means that all of the material must be published before 1923 or unpublished and created by persons who died prior to 1940.5 But mixed collections of 20th-century material will also contain public-domain materials, although determining exactly which items are which may be difficult. Publication date and creator’s death date are usually discoverable what is more difficult is the determination of the copyright status of materials published between 1923 and 1989. John Wilkin has written a comprehensive analysis of this problem, which he calls “bibliographic indeterminacy.”6 Indeterminacy is only a problem, however, if one is seeking certainty. From the perspective of risk management, it is enough to recognize that most collections of 20th-century materials will contain some public-domain materials. Some materials, for example, will have been published between 1923 and 1963 without copyright notice, or will not have had their copyrights renewed. This category of works is estimated to include about 55% of the books published during this period,7 and it will often be true of newspaper or magazine clippings as well. Another group of materials will be works of the US federal government, which are not eligible for copyright protection. Amongst the unpublished works in a given collection, such as letters, some will have been created by people who have been dead more than 70 years. Once a risk-management approach is adopted, these categories can be seen as broad groupings that reduce the number of risky items in a potential digital collection. Even though exact determinations cannot be made, recognition of the categories and their potential application is an important step in deciding with some accuracy how risky or safe a particular proposal may be. The second strategy for risk management in digitization projects is to ask permission from the people or organization that would be most likely to object to the digital display. Again, we should recognize that it will usually be impossible or impractical to identify every rights holder and ask permission, and no project need depend on meeting such an impossible standard. The principle of reducing the number of people likely to sue suggests that asking permission even from only a few rights holders, especially those who seem likely to hold rights in a substantial portion of the included material, is an important step in risk management. If there is a large number of clippings from a particular newspaper whose publisher still exists, or a large number of letters by a single author whose heirs can be identified, these are good candidates for permission. Literary estates, which often police the use of works by a particular author, are also good candidates for permission. But it cannot be emphasized too often that asking permission from some large or prominent rights holders does not mean that permission must be obtained for every item in a digital collection. The goal is to reduce the number of likely plaintiffs and to head off those who seem most likely to object as part of an overall risk-management strategy. This approach can significantly increase confidence without creating an insurmountable obstacle. Also stemming from the principle of reducing potential plaintiffs is a third strategy of having, in advance, a take-down policy for any materials made subject of a complaint. Digital collections generally