Infrapuncture is a comely term at a moment in time when there is a lot of discussion around the political roles of automated agents in digital infrastructures. Before we move any further, perhaps a short definition of bots is necessary. When we say bots, we refer to code that automatises certain behaviours and which often acts as an interface between platform and human users.
Many online communities engage with bots, for example the editor community of English Wikipedia, which consists of both humans and bots. The interactions between them go beyond the maintenance of Wikipedia. Instead, affective relations are formed wherein the bots are anthropomorphised. So writing a bot implies not only to understand the API (Application Programming Inferface) of the platform, what determines the possibilities of interaction, but also the social norms established within the community of editors and users of Wikipedia.
And that's of course just one example. Bots act differently depending on the platform on which they are running.
## Bot logic
What kind of puncturing logics might bots enable in digital platforms?
Bot logic is phrased as a response to platform logic, which Jonas Andersson Schwarz describes as "digital platforms enacting a twofold logic of micro-level technocentric control and macro-level geopolitical domination, while at the same time having a range of generative outcomes, arising between these two levels".
'Bot logic' refers to the situational effect of bots upon a socio-technical ecology and their potential to infiltrate and co-exist with server-side conditions. Perhaps we can move our attention to these few points. When referring to platform logic in the points below, we refer to commercial infrastructures, not federated and free software platforms such as those present in the Fediverse, which have a different kind of dynamics[^theses].
* Where platform logic accumulates, bot logic disperses
* On commercial platforms, the engagement of users equals economic value that is translated through data capture and organisation. Metadata is extracted from users that then through pattern matching can be used to target users for advertisements. While bots can and do participate in this economy, they can also enable its sabotage. In the case of buying bot followers, this can be a means to generate noise in the collected dataset and blur the perception of the user as a set of behaviours that the platform has.
* Where platform logic centralises, bot logic fragments
* Platforms such as Twitter or Facebook use a centralised system, in the sense that the servers on which our information is stored are owned by their company. Bots, on the other hand, do not require a lot of computational power in order to run. They can be simply run from the computers of the persons who wrote the code themselves. In fact, bots really point to the materiality of the structures on which they run, as researcher Stuart Geiger also points out when he talks about 'bespoke code' as code that extends or transforms the operations of software platforms, but "runs on top of or alongside existing systems instead of being more directly integrated into and run on software-side codebases".
* Where platform logic creates distance between user and infrastructure, bot logic develops an intimate knowledge of the platform
* If we consider means of communication as means of production (Williams, 2005), there is a certain alienation that happens on commercial centralised platforms, where the user has no stake in the development of the material conditions of the platform on which they communicate. From this point of view, the making of bots implies a closeness to the platform that is indicated through the understanding of both the sociological and technical systems that determine the usership of a platform. In order to code a bot, you need to know what kind of actions are allowed and how the bot would be received by the community.
* Where platform logic reinforces habitual behaviour, bot logic encourages new habit formation
* If we think about a commercial platform as a structure or surface on which actions can take place, these actions are often predefined by the affordances of the platform. However, as was mentioned in the beginning, bots are the automation of certain actions and behaviours. To be able to define these behaviours as a user can mean an alteration of the socialities embedded in a platform.
All of these points were written with commercial platforms in mind, however, exciting developments are happening in federated platforms such as Mastodon, where users are part of defining features and possibilities of interaction. There, the norms of the platform and the way they are codified into the technical structure are more often revised and reformulated together with the user base. This in itself creates a different space for bots, which are still active contributors in the way sociality is imagined on these platforms. However, on platforms like Mastodon, bots don't only comply to the terms of services of the API but also to the code of conduct, for example.
