Who do you Trust? The Authentication Problem
How do we know when a digital document is "authentic"? While many in the library and academic communities hope that there will be a technological solution, the reality is that technology alone cannot solve the problem of authenticity. A report this week of research at a Chinese university illuminates one reason for this: technical tools are subject to failure, compromise, forgery, and hacking.
- U.S. mulls new digital-signature standard, By Anne Broache, and Declan McCullagh, CNET News.com, November 1, 2005.
The article reports a flaw in an official federal standard that was originally devised by the National Security Agency and is widely used to create and verify digital signatures in e-mail and on the Web. In fact, it is embedded in every modern Web browser and operating system. The CNET article notes that, while the flaw that Chinese scientists discovered in the "Secure Hash Algorithm" is "theoretical," it will eventually make it easier to forge electronic signatures.
But authenticity requires more than secure software. Even if we had a tool that could never be hacked and that would last forever, we would still only have part of a solution: the technical part. The other part of the solution is social: it is the issue of Trust.
Software provides the technical part of the solution
The technology of authentication provides a way to verify that a document is what it purports to be and determine if it has been altered or not. Document-creators can use software to create special files (called "hashes" or "signatures" or "keys") based on the original document. These special files are typically stored with a "trusted third party" -- neither the document creator nor the recipient. Document-users can then use software to check the authenticity of the document in hand against that "hash." The software is able to determine only if the document in hand is identical to the original. Even the smallest change (e.g., the insertion or removal of a blank space) will result in a report that the documents are not identical.
Trust is the social part of the solution
But this technological check does not solve the authentication problem by itself. The check against the hash is only as reliable as the trusted third party. The software just gives us a technical means of shifting who we trust -- instead of trusting the party that delivered the document to us, for example, we trust a third party that tells us that the hash is correct and authentic. If the hash isn't authentic and unchanged, the check against the hash is worthless.
This concept of a trusted third party is, therefore, an essential component of the authentication chain. That should lead us to an important question: who will we choose as our trusted third parties? This is important because the tools only work if we can trust the third party to do its job. In the case of government information essential to our democracy, this trust has to last forever.
Who do you trust?
Ask yourself who in society is the most trusted third party in delivering information? The government? The press? Publishers? Technology companies like Microsoft and Verizon?
What about libraries?
Now ask yourself what we will do if we think that technological-verification is all we need to ensure authentication and we find one day that the tools have failed as described in the CNET article.
A Social Solution built on Trusted Institutions and Legal Deposit
Trust is a social phenomenon, not a technical one. What if, instead of putting all our faith in potential technological "solution" for ensuring authenticity of government documents, we instead relied on the existing infrastructure of depository libraries to ensure authenticity through their collective possession of multiple copies of digital government publications, distributed by GPO at the time of their publication under the legal-mandate of 44 USC?
This solution promises to be a sound, sustainable one because it relies on libraries as the trusted repository of information. Libraries have a long, well-established social role of providing information; people trust libraries because of it. Libraries have a vested interest in ensuring that the information they provide is authentic and people trust them to do so because it is their primary mission -- not a byproduct of publishing or making money or the various missions of government agencies.
The trust people place in libraries in general can be increased in the digital environment by relying, not on one or two libraries, but on many libraries with different funding streams and missions. Any unforeseen compromise in one institution becomes a single error in a large system of information-provision. (See Article outlines bottom-up standards for digital preservation systems.) Even in the paper and ink world, forgeries are possible -- though more difficult than in the digital world -- and one important way we determine authenticity is by comparing multiple copies.
A different approach
This approach is subtly different from the approach of hoping for a technological solution to authenticity. It recognizes that the social issue of trust (along with the existence of multiple copies controlled by different parties) is paramount and the role of technology is secondary. The role of technology is simply to provide tools to help implement that trust. Indeed, if we used this social-trust legal-digital-deposit approach, libraries would still use technical tools (e.g., LOCKSS, PKI, state of the art hash technologies) to validate the integrity of digital files. Combine these tools with trusted institutions, legal deposit, and multiple copies under multiple jurisdictions and you have fail-safe a recipe for ensuring authenticity.
The problem with hoping for a technological solution was clearly articulated back in 2000 by Abby Smith, Director of Programs at the Council on Library and Information Resources.
Interestingly, the scholar-participants suggested that technological solutions to the problem [of establishing the authenticity of a digital object] will probably emerge that would obviate the need for trusted third parties. Such solutions may include, for example, embedding texts, documents, images, and the like with various warrants (e.g., time stamps, encryption, digital signatures, and watermarks). The technologists replied with skepticism, saying that there is no technological solution that does not itself involve the transfer of trust to a third party. Encryption -- for example, public key infrastructure (PKI) -- and digital signatures are simply means of transferring risk to a trusted third party. Those technological solutions are as weak or as strong as the trusted third party. To devise technical solutions to what is, in their view, essentially a social challenge is to engender an "arms race" among hackers and their police.
-- Digital Authenticity in Perspective in "Authenticity in a Digital Environment," Council on Library and Information Resources, Publication 92. (May 2000).
James A. Jacobs, November 3, 2005