Hidden algorithms and AI challenges towards historians

 When I am asked to define history as a subject I often find myself in a complex situation. Since atleast the 1960's history have been influenced by various sciences, from sociology to political science and urban planning. History thus have not a theoretical foundation, instead we would speak of multiple histories. However, one aspect that I often highlight is that history have a large focus on contuinity (what remains the same) and change as well as explaining how the human experience are subject to both genealogies and sudden breaks. The second aspect I often speak about is that history have a very strong source criticism, an aspect that comes from the launch of history as a academic subject in the 19th century, when Leopold von Ranke started to write history from actual archive records rather than myths and folklore. This aspect of source-criticism is crucial to both my identity as a researcher and as an historian. However, given the rise of digital society, the possibility for source-criticism is meeting some challenges. 

In around two decades social media will come to play an important part of the history writing of the period from the end of the 1990's and onwards. Social historians will find themselves grappling with questions such as how did the black lives matters-spread, and which deep culturally rooted assumption did it challenge by it's online mobilization. This is however a research question that also will demand an indepth knowledge of technologies due to two factors. The first factor is that the technologies themselves both limitis and creates the possibility for historical actors to act. In this regard, social media can create a basic infrastructure for political actions but which political actions that becomes possible are regulated by Facebook:s algorithm. This have for instance been highlighted in a study from 2017 by Kaun & Uldam which showed that the most frequent viewed posts in a Facebook-group where not the one most needed to be spread by the activist collective, but rather the ones with most likes. Hence, the actions that actually are of importance to social movements will probably not be seen in a pure reading of Facebook posts. This is also complicated by the fact that Facebook's algorithms which manage the spread of information is a business secret, thus making it hard for researchers to fully grasp the context in which continuity and change occured.

Another aspect has to do with the notion of transparency in qualitative research. Since the humanities to a large deal works with subjective interpretations of texts, transparency is of great importance. In conventional humanities transparency is showned by discussion of both the method and theory, which is the foundation for studies, but what happens when historians will become more and more dependent on AI for doing their research? On the one hand, AI can for instance be of great use when reading through hard to read texts such as ancient manuscript written by scribes. Here AI can be utilized for data-mining, where it both can transcribe the manuscrips and investigate relations between historical actors. But since AI operates from a different point of depature than humans, it will be hard to learn how it actually reaches it's conclusions. To fully discuss this, historians need to have technological competencies which right are not part of classical humanistic training and probably never will be due to the amount of competences needed to have a meaningful conversation on the technology. 

The world has come a long way since the first
desktops (Picture: Wikipedia)


In my view, AI and social media thus comes with many potential benefits for future generations of historians but also with some limitations. I also think that we are just in the beginning of the process to fully understand what the digital revolution will bring about. I am however sure of two things where the first is that history as a subject will continue to be crucial in contemporary debate, since history is our main source of knowledge. I also think that in time, such challenges will be met but the only question is how, I guess time will tell. 

Comments

Popular posts from this blog

The use of history and why historians should mind the gap

Contemporary Sexual Politics: Efforts to Silence Sexuality in Politics

Contemporary Sexual Politics: a Background