Bots as Digital Infrapunctures

We have traced the orientations of digital infrastructures and computational infrastructures in previous tracks. However, bots are often contextualised as acting on platforms. What is the difference between these terms and where do they overlap?

In many ways infrastructures and platforms overlap in their invisibility, broad public usage, or extensibility. According to Plantin et al (2016), both ways of framing offer helpful elements for their analysis. We are witnessing a platformisation of infrastructure in tandem with an infrastructuralisation of platforms through information technologies. Here we find, on the one hand, infrastructures start to splinter into services taken over by private enterprises, and on the other hand, platforms start taking on more responsibilities which were previously managed by the government¹.

For the purposes of this online module, we are interested in the programmability (what can be build on top of the offered functionalities) mentioned in the previous track and affordances (what is made possible through a design) of platforms combined with the valorisation of public interest and accountability systems that are characteristic of infrastructures (through standards and public funding). However, in order to highlight the importance of optimization practices for a public interest, and not for corporate profit, we will from now on refer to digital infrastructures. Doing so avoids the ambiguity of describing the activity of repair for different kinds of interest, which could include corporate interest. We are interested in the potential of bots to repair in the benefit of one or multiple public interests. However, as we stay close to Deb Verhoeven's terminology, we will not forget the shaping forces described by Seda Gürses.

As programmable objects bots have particular action modes.

Below some examples (although this is a non-exhaustive list).

• repetition: bots can be run repeatedly,
• condition: bots are often written in response to a particular condition,
• iteration: bots can be used multiple times,
• memory: bots can rely on a database,
• tempo: bots can operate at a specific time frame as described by the programmer,
• versioning: bots can make multiple versions of an original source,
• amplification: bots can extend the reach of a message.

Of course, these action modes can also be executed by people.

It is by no means surprising that many Twitter users are mistaken for bots, or that the term itself has attained a derogatory meaning. Human users might be called a bot as form of insult. However, an interesting phenomenon can be observed on digital infrastructures such as Twitter, where human users automatise certain actions similarly to bots to create networks of dissent and to push activist anti-racist, feminist, queer counter-narratives.

Such a moment happened recently on Dutch Twitter. In response to the Black Lives Matter protests, extreme right wing politician Geert Wilders posted an image on Twitter on June 5th 2020 using the hashtag #ZwartePietMatters.² Following this post, a wave of fancam³ users from the k-pop community flooded the hashtag with video recordings of their favourite k-pop stars, making the thread difficult to follow. Such practices are becoming a common phenomenon across the Twitterscape, where fancams are used not only for praising musicians, but instead for derailing and hijacking hashtags these users consider unacceptable.

Fancam users display an intimate understanding of the digital infrastructure, so much so that they manually generate noise through collective interventions in various topics. The repetitiveness of their attacks and the high amount thereof clog the thread interface, thus altering its functionality. In this sense, human user interventions can also be phrased as bot logic being applied.

Footnotes