By Cara List
The Museum Education Site Licensing Project (MESL) presents museums and institutions of higher learning with a unique opportunity to explore the educational potential of making museum images and related textual material available on computer networks. Sponsored jointly by the Getty Art History Information Program (AHIP) and MUSE Educational Media, the project will act as a laboratory for the distribution and use of a digital image database. Participating institutions hope to study and find solutions to problems currently inhibiting the development of computer based learning tools. The developers of the project see among their primary goals the resolution of issues of intellectual property rights and network security. In making digital access to museums' collections available on university campuses, these issues can be practically examined and guidelines for future collaborations may be established.
MESL is a limited engagement, running for two years, from September of 1995 through the end of the 1996-97 academic year. Seven museums of art and culture, including the Fowler Museum of Cultural History at the University of California, Los Angeles; George Eastman House; Harvard University Art Museums; the Library of Congress; the Museum of Fine Arts, Houston; the National Gallery of Art; and the National Museum of American Art, will make a minimum of 7000 images and related descriptive text available over the course of the two years. This database will be mounted on Web sites and be made available to the campus network users of seven universities: American University, Columbia University, Cornell University, the University of Illinois at Urbana-Champaign, the University of Maryland, the University of Michigan, and the University of Virginia.
In creating the framework for MESL, it's organizers have devised a laboratory where they will be able to study and resolve many tricky issues that would certainly trouble any more permanent site licensing agreement of this nature. Paramount among these issues is the potential for infringement of copyright law, primarily as it applies to the museum images that are the distinguishing ingredient of this database. Copyright law is not a cut and dried area of the law to begin with, and when brought to bear upon visual art, it is even hazier.
Copyright Law deals with intellectual property or the intangible, non-physical element of content as opposed to body. In this way copyright law is different from property law, which governs physical property. The concept of intellectual property is much easier to grasp as it pertains to literature, music and many of the other products covered by copyright law, as they are more frequently mass produced objects, and their physical presence functions only as a medium by which their real value is conveyed. People do not usually consider the intangible element of fine art, which, like illustrations, maps, and other mass produced images can be reproduced. By it's definition a work of visual art is a unique object, and valuable for itself.1 We tend to think of the content and the body of fine art as unalterably tied to one and other.
Under the 1976 act however, it is clear that the copyright to an art work is distinct from the ownership of that work unless it was a work for hire. The artist retains his or her copyrights unless a transfer of copyright ownership is in writing, and signed by the artist.2 While manuscripts, scores of music, dramatic works, and other copyrightable material ordinarily go through a publishing process, complete with a transfer of copyrights, visual art does not. Many works are sold without any transfer of copyright.
For most of the images and objects which are being used in the MESL database, copyright ownership is not an issue, as the art works are older, and the copyright lies in the public domain. However, the museums who own these objects wish to control the use of them, profiting themselves from any reproduction. When the museum makes a reproduction of an art work in their collection they have created a derivative work,3 which they then hold the copyright to. The museum will also quite likely disallow photography within their walls, thus holding the corner on the market for the reproduction and distribution of that image. In a conversation with Susan Kornfield, a prominent copyright lawyer, she expressed to me her opinion that this interpretation of the term derivative work is a misreading of the law. In order for the slide to be a derivative work, some additional authorship would have to be added to the image. With consideration to MESL and any similar project, this becomes a moot point. The museum must be confident in the belief that it is the owner of the intellectual property, and as such, is able to take steps to protect the right to benefit from it's use. The museum would not risk making the image available to the public in a situation where it could not exert complete control if there was any risk of loosing potential profits.
The whole issue of copyright becomes both more difficult, but more clear cut in the case of contemporary art. Occasionally when works of art are bought by museums, contracts are also drawn up for the transfer of rights to reproduce, distribute, and display. Awareness of the importance of such understandings is becoming more common among artists as court cases such as Richard Serra's Tilted Arc case4 and Donald Judd's against the Whitney Museum of American Art5 are publicized. The Visual Arts Act6 of 1990 has also served to bring artists up to speed in protecting their rights. However, occasionally, as with a young or unknown writer, an artist will sign away his or her copyrights to make the sale or gain exposure. In these cases, the museum's position is clear as to their ownership of copyright.
There are essentially only two possibilities open to the museum when dealing with an image for which it clearly does not own the copyright. The first is an obvious solution, but not always the best in light of the educational purpose of the project; to not include the piece in question in the database. However, the selection process involved not only the museums but the universities as well. At the February 7-9, 1995 meeting of MESL participants, it was agreed that in addition to lists of available works from the museums, selection would be based on requests for specific works from the universities.7 The second option then, was for the museums to clear reproduction and distribution with the artist. In many cases this was not possible, but the Eastman House was willing to explore the option in order to make contemporary photography available and the National Gallery of Art has already cleared some works of art for use.8
Assuming that the museum is the copyright holder, and is free to utilize the image in the MESL database, what are the issues of copyright that must be addressed? In making the images available to a large, albeit, limited user group, there are many potential infringements which must be guarded against. In the Museum Educational Site Licensing Project Cooperative Agreement, certain responsibilities are outlined that give some framework to what is and is not possible:
Each institution agrees to hold the other participating institutions harmless from claims or liability arising from the use of images and information in the project. Contributing institutions agree that during the project the images and accompanying information may be used at other participating institutions for educational purposes, including faculty research, teaching, or student projects. All participating institutions agree that neither the images nor accompanying information will be used for non-educational or commercial purposes, or redistributed for any purpose beyond the participating institutions without the prior written permission of the contributing institution. Before making any further distribution beyond the participating institutions, or engaging in any non-educational or commercial use of the materials, users must seek and obtain the written permission of the contributing institution.
Participating institutions who make the images and accompanying information available to their authorized users agree to inform these users in a manner agreed upon by the project participants of the proprietary nature of the material and the limitations on its use. (P. 3)
In fulfilling the final issue listed above, the MESL Welcome page includes hypertext that links to their "Conditions of Use" site. The rules that the users of the images and data must adhere to are clearly stated on the "Conditions of Use" Web page.
In framing their agreement for what is allowable within this experimental project, the participants took a deliberately liberal stance. Each institution entered into the collaboration in good faith, with the interests of education and experimentation in mind. Many choices have been left to the future, to be resolved when it becomes apparent what the perimeters of the problems are. A large portion of the discussion about intellectual property issues has been speculative to this point. At the February meeting of project participants, the group contemplated some of the potential challenges that they would need to address in the course of the project. Among some of the more extreme examples of situations where copyright issues will come into play are: repackaging of the material for mastering internal CDs, designing a multimedia course or creating an electronic textbook; use of the material in visual thesauri; use of the images or data by quasi-commercial enterprises such as the University Press; use of the material by collaborators who aren't part of the MESL consortium; contributions of annotations based on collaborative research done at universities; and manipulation of the images for study purposes.9
At the June 5-7, 1995 meeting at the University of Maryland, solutions to some problems were discussed. It was agreed that off loading of data to floppy disks and CD-ROMs by faculty will be allowable if the material is used for educational purposes, but that any further distribution or use will require negotiation. It was also agreed that the universities will carry the burden of identifying any group of users outside the immediate university community. For instance, accounts are offered by the University of Maryland to K-12 groups, and by the University of Michigan to alumni. Specific permission will be granted to these groups upon application for access.10
Some issues such as encryption for network security have been worked out with the universities on an individual basis, depending on what systems the schools already had in place. Extensive discussion has taken place over the relative merits of various methods of encryptions such as the Kerberos system utilized at the University of Michigan.11
Currently at the University of Michigan encryption is not being employed to protect the data from the MESL web site. John Weise, the Systems Research Programmer who is in charge of the technical task of mounting MESL on the University of Michigan network, explained that they want to (at least) start out with limited security measures so that an optimal number of users will have access to MESL. In this way, the organizers of the project will be able to determine how and by whom this database is used. MESL is not, however, available to just anyone on the web. The University of Michigan's server software has the ability to recognize the domain name of the user, and restrict access to umich.edu users only.
It is possible in the future (almost certainly if the project is carried beyond it's experimental stage) that some form of increased security will be employed. One possibility suggested by John Weise would be to utilize one of the servers offered by Netscape. The Netscape Commerce Server would allow the University to publish the MESL database on the web with advanced security features such as encryption and authentication.12
Encryption is a two stage method of data scrambling. Before information is transmitted over a network the data is scrambled or encrypted, to avoid trespassing by unauthorized users. The second part of the process is authentication. The user who has requested the data is identified, and authenticated for access to the data. When the data arrives at it's authenticated destination the data is unscrambled so that it can be read. A network such as a University system might opt to use various encryption techniques to ensure that data in e-mail, course folders, or web sites could only be read by the appropriate users or user groups within it's domain.
There are essentially two types of data encryption schemes. A symmetrical or private key system requires both the user and the sender to employ a private, or secret key to encrypt and decrypt the information. A public key system works with two types of keys: a public and a private key which work together to encrypt and decrypt data. This system operates in two ways: one verifying the sender, the other the user. To verify the user, information is sent out employing the user's public key. When the information arrives at it's destination, the user's private key is the only key able to unscramble and read the data. The Netscape Commerce Server employs public key cryptography from RSA Data Security Inc. where public key cryptographic technology was developed.
The University of Michigan currently utilizes Kerberos, a private key technology for some of their network functions. The Kerberos system is simpler and requires less processing power. When the user logs onto the network using his or her Kerberos password, the uniqname and password are encrypted into that computer with a private key. The server controls a directory of all the private keys of it's authorized users. The login data is sent to the server which compares it with the directory list, authenticates the user, and then determines whether the user is permitted the service being requested.13 At the University of Michigan, WWW browsers are currently unable to operate in a Kerberos environment, and modification would be necessary to implement such a system.14
Encryption may in fact be an extreme measure. None of the supplying institutions involved are as interested in limiting access as they are in controlling the use of the images. They have expressed a great deal of concern that their images don't resurface in commercial uses in the future. To this point the issue of resolution has been debated at length, with some misunderstanding between universities and museums as to what the meaning of "high resolution" is. Museums, whose primary concern is with the ability of users to make high quality prints, were unaware that the high resolution that universities were discussing would be suitable for digital transmission, but would not nearly approach the requirements for color separation. It was agreed that if users wished to utilize higher resolutions for concrete applications, that specific permission would be required, and tracking of the use would be employed.15 However there are many digital uses such as multi-media production that could utilize the images at the resolution at which they are available.
The museums are quite reasonably interested in being able to track and control the ways in which their intellectual property is used. Several potential solutions to this problem are watermarking, fingerprinting, and digital signatures. All three are technologies making it possible to embed information in the image, however, they differ significantly in their technique.
Watermarking refers to the technique of stamping paper with an identifying mark that is invisible until held up to the light. The most common form of digital watermarking can be seen any time you tune into cable television. Surfing through the channels one might see a small semi-transparent channel identifier in the lower right hand corner of the screen, signifying that the program belongs to MTV, CNN, TBS or one of the many other operators. The same technology can be applied to images transmitted over the World Wide Web. The unfortunate aspect of this form of identification is that it is, or can be obtrusive. In the case of art, it may hide an element of a painting that a researcher wishes to study. Another problem with this technique is that it can be tampered with: airbrushed out, cropped off, or in some other way, manipulated out of existence.
Watermarking has become a somewhat generic term, also covering two other techniques employed to embed data in images, video, or music. One method is being developed by a small British software company, MOR Ltd. The technique known as FBI, or Fingerprinted Bitmapped Images was invented by Finnish programmer Otso Paatelma. The code for the fingerprint is scattered throughout the image, and is invisible to the human eye. The system functions by altering individual bits in a series of pixels to store the fingerprint which can be an alpha-numeric sequence such as a name or other copyright ownership information. The changes take place in the least significant bit in the byte, and shifts the brightness or color so minutely that even with high magnification, changes are virtually undetectable. Because the fingerprint is scattered throughout the image, it is impossible to destroy it through cropping or digital manipulation, and because it is invisible, it can't be found to be altered. There are, however, two major problems with FBI. It can't be easily decoded, and the computer file storing the stolen image would have to be accessed before infringement could be proved. This problem also means that a well meaning "borrower" might not be aware of the fact that he or she was infringing on the copyrights of another. The second, and more serious problem is that FBI cannot survive los-sy compression methods. The least significant bit is terminal in a compression technique such as j-peg, and the fingerprint disappears.16
The most promising of the new watermarking techniques is the product of a Portland, Oregon company called Digimarc Corporation. Their technique which they call a digital signature is similar to FBI in that the signature is scattered throughout the image and is invisible to the human eye, but is considerably different in it's effectiveness. Digimarc's digital signature can be embedded with two kinds of code patterns. A public code pattern, which is based on an industry wide standard, can be deciphered and read by any user. It could contain information about ownership and licensing, and who to contact for permission. The image would also have other information embedded in it with a private code pattern. This code would be known only by the creator/owner and would be unique. It could contain other identifying information, and would be irrevocably embedded. Because this code is private, it could not be found and tampered with by anyone with less than honorable intentions. If the owner chose, he or she could choose a different code pattern for each released copy of the image, so that if it turned up gracing the side of a coffee cup, the origination point could be found.
The most appealing aspect of the Digimarc system is it's robustness. It is holographic, meaning that all of the signature is embedded in one place. It does not rely on the "least significant bit" technique; instead, it masquerades as random digital noise. Noise refers to the unavoidable, tiny, random variations in an electronic signal.17 Because of this, a larger image can contain more information while retaining it's robust quality than can a very small image. Most notably, this technique also provides the signature with the ability to survive multiple generations of copying, printing, scanning, and los-sy compression. It also provides the copyright owner with the ability to find their code in a possibly stolen image by scanning that image and using their private code pattern to decipher their digital signature.18
For the organizers of the Museum Education Site Licensing Program, the issues of encryption, authentication, watermarks, fingerprints and digital signatures remain unresolved, and in the still distant future.19 Because this project is a work in progress, a great many issues remain to be resolved as the database is implemented and used. The institutions involved have entered into this project in a cooperative framework. They are excited about the educational potential, and want to work together to make site licensing a viable venue for museum databases in the future. While the museums do not want to find their images showing up on screen savers, their real concern is with furthering education and access to information on the arts and culture. With that in mind, they are hesitant to overburden MESL with heavy restrictions. Above all, it must be remembered that this is an experiment. There will surely be successes and failures, but without projects such as MESL, no progress can be made. Karl Katz, the director of MUSE Educational Media expressed this view, "We will never achieve the 'virtual museum' of the future without a common framework of rights, permissions, and restrictions. The collaborative nature of the Museum Educational Site Licensing Project will ensure that the solutions it develops will meet the needs of both the museum and the educational communities, bridging the gap between image users and rights holders."20
Crawford, Tad. Legal Guide for the Visual Artist. New York: Madison Square Press, Inc., 1987.
PAPERS AND ARTICLES
Arthur, Charles. "Digital Fingerprints Protect Artwork." New Scientist 144 (12 November 1994): 24.
Borsook, Paulina. "Highway Safety; The Key Is Encryption." Byte 19 (March 1994): 60.
Caruso, Denise. "Digital Commerce: Two Plans For Watermarks, Which Can Bind Proof of Authorship to Electronic Works." New York Times, 7 August 1995. New York late edition.
Fitzgerald, Mark. "Invisible Digital Copyright ID: British Company plans to Soon Formally Introduce the Technique." Editor and Publisher, The Fourth Estate 127 (25 June 1994): 62.
Intellectual Property and the National Information Infrastructure. "The Report of the Working Group on Intellectual Property Rights." Dept. of Commerce, Washington, DC, September 1995.
The J. Paul Getty Trust. "Pilot Project to Explore Intellectual Property Rights and Other Issues Related to Digital Images and Information." Press Information. Santa Monica, CA, December 23, 1994.
Museum Education Site Licensing Project. "Cooperative Agreement". June 7, 1995.
Museum Education Site Licensing Project. "Report of Participant's Meeting." Smithsonian Institution, Washington D.C., February 7-9, 1995.
Museum Education Site Licensing Project. "Report of Participant's Meeting." The University of Maryland, College Park, MD, June 5-7, 1995.
Robinson, Walter. "Tilted Arc Goes to Court." Art in America 75 (February 1987): 168.
Scott, Karyl. "Encryption Schemes Put Safety First: Vendors Are Providing Kerberos and RSA as Tried-and-True Ways of Securing Data." Data Communications 20 (21 March 1991): 17-18+.
"A Cure for the Common Code." The Economist 320 (24 September 1991):104+.
"Digital Museums." Technology Review 97 (November/December 1994): 20.
"The Rights Stuff: Buying and Selling Art in a Digital World." Scientific American 272 (January 1995): 30-31.
WORLD WIDE WEB SITES
"Conditions of Use." World Wide Web http://mesl.itd.umich.edu/conditions.html.
"Digimarc Corp. Announces New Copyright Protection Technology." World Wide Web http://www.teleport.com/~digimarc/annon.html.
"Frequently Asked Questions About Digimarc Signature Technology." World Wide Webhttp://www.teleport.com/~digimarc/faq.html.
"Kerberos Authentication." World Wide Web http://hoohoo.ncsa.uiuc.edu/beta-1.5/howto/kerberos.html.
"The Museum Education Site Licensing Project"World Wide Web http://mesl.itd.umich.edu.
"Netscape Commerce Server." World Wide Web http://home.netscape.com/comprod/netscape_commerce.html. 1995.