Algorithms are often referred to as ‘rules’, computationally applied to solve an array of mathematical problems. In fact, this limited definition belies the complex role that algorithms play in human society. Everyday we experience non-computational algorithms and we adapt our human behaviours accordingly. For example, joining a queue at the supermarket instead of going straight to the front is an example of human integration with an ancient non-computational algorithm: the practice of spatially ranking ourselves according to the time we arrive to join a queue. We’d all like to be at the front, but deference to the protocol of ‘lining-up’ behind those who arrived earlier, dictates we wait our turn. Before we can consider what algorithms achieve, what they do, and how they make things work in computational networks, we need to first grasp the point that algorithms are a historical product of pre-computational thinking. Humans have been using algorithms of one kind or another for thousands of years to sort and create order out of chaos, to select who goes first and when this occurs. Needless to say, supermarket shopping would be a much more challenging experience if a differently applied rule mandated we all rushed to the front at once and demanded service.
Algorithms in computational networks, similarly, operate alongside human behaviours to generate an orderly result, often related to some form of sorting, ranking, sifting or filtering. They must be well designed to avoid chaos, but not so strictly structured that they engender autocracy. Critically, in this current climate of machine learning and the new artificial intelligence (AI), algorithms are no longer pre-programmed mechanisms: increasingly, algorithmic agents are adaptive and changing according to the feedback they receive from human actors in the field to which they are applied. Due to their new pairing with machine learning, algorithms are trained to adapt to data inputs such as advertising choices an Instagram user makes or which Facebook friends we click on the most in order to decide whose future posts we see first. Taking this perspective, in a forthcoming article I will be arguing that we might need to re-consider algorithms in the role of nonhuman participants in a digital network, rather than as simple structural protocols. Such a consideration gestures toward an expanded understanding of automated data networks as intrinsically linked to human behaviours and practices.
In such a scheme, the ‘digital’ is further engaged as a kind of liquid current within which we swim everyday. Some of the practices I will be examining, flowing through this current, concern blockchain networks and their relation to the ‘natural’ environment, creative labour practices of Jamaican music producers via platforms like Instagram, and the far from seamless sensory interactions between humans and computer vision technologies such as AR and VR.
At a time of extreme and accelerating climate change, technical networks such as those increasingly governed by algorithms, provide the essential ground for human linkages with our planetary environment and with nonhuman others sharing the same habitat. It is essential that we no longer examine technical networks as separate from human relations. Rather, the imperative is to examine the techniques and methods that connect digital (and algorithmic) forces to human cultures and social assemblages.