BloggERS!, the blog of the Society of American Archivists’ Electronic Records Section, recently published “Mecha-Archivists Revisited: An Interview with Trevor Owens and Emily Reynolds.” Owens and Reynolds (both IMLS) discussed the IMLS’ national digital platform funding priority, its impact on the work of digital archivists, exciting projects coming out of the funding priority, and the concept of mecha-archivists.
BloggERS!: Trevor, in a 2014 blogpost, Mecha-Archivists: Envisioning the Role of Software in the Future of Archives, you highlighted the potential value of computational techniques, such as topic modeling and named entity recognition, to help “extend and amplify the seasoned judgement, ethics, wisdom, and expertise [of archivists]” to support making materials available to the user. What progress have you seen in this space since 2014, and how would you rate the development of and training around these tools as a priority within IMLS?
Trevor: I see a lot of the ideas I explored in that Radcliffe Workshop on Technology and Archival Processing as fitting very well with the idea of the National Digital Platform. The key concept in that talk is that we need to approach the work of cultural heritage institutions as complex systems which deploy enabling technologies that support, amplify and extend the abilities of archivists, librarians and the curators to do their work. I realize that’s a mouthful. So I can talk through some examples.
All too often, I have seen folks approach some computational tool and say, “Oh, we could use this to automate classification or description” or a variety of other activities. This sets the bar way too high for the machines. It also is part of longstanding, problematic and flawed notions about expertise, efficiency and labor that devalue what it means to be a professional and an expert. The judgement of professionals and experts is really hard to beat, and it isn’t something we should be trying to beat. Instead of erasing or ignoring all of the accumulated wisdom and expertise of professional librarians, archivists and curators, we should be working to build from and amplify it.
In my mind, the solution is rather simple. Instead of trying to replace the work of experts, it is much better to think through how we can enable and extend that judgement through tools. The example I used in the Radcliffe talk involved Topic Modeling, but I think the same process can and should work for things like natural language processing tools named entity extraction tools, or for that matter tools and services that automate deriving data about audio files or image files.
I see all of this fitting into the national digital platform in a few clear ways. First off, the platform is defined not as a set of tools and services, but as the combined effects of those tools and services and the professionals that animate and operate them. In that vein, the platform is as much about empowering, training and supporting professionals to do the work as it is about giving them the tools to enable them to do the work.