• No results found

4 A Politics for the Insensible

4.2 Algorithmic animism

4.2.3 Non-human affinities

References to imperfect, unpredictable, potentially corrupt ‘mushy humans’ resonate from Bitcoin throughout aspects of Ethereum, as a project to realise a general platform whereby no human would be trusted or able to control it. It gave rise to a very particular notion of autonomy of the system, expanded through Smart Contracts to ideas of the DAO. Such ambitions are made to matter by for example making contract accounts on the Ethereum network able to ‘call’ functions in other contract accounts without the need for human prompting. As clusters of Smart Contracts and DAOs expand and interact with each other, systems of automated processes might set in motion in ways that can potentially develop way beyond immediate oversight and understanding of humans, and in ways that, in theory, would be unstoppable. These ambitions also provoked speculation about what might happen when cryptographically unbreakable, self-executing blockchain systems might be wielded by a future artificial intelligence. From the very beginning of Ethereum there was therefore a giddy self-conscious excitement about possibly creating an unstoppable algorithmic authority, questioning whether the platform represented a ‘freenet or Skynet’ (Filippi, 2014).85 So far, actual implementations have turned out to be more mundane, looking towards efficiency gains of automating aspects of contracts and payments, and not quite able to fully sever the ties to the mush the insensible and indeterminate. Yet there are tendencies in the community that actively strive towards such a realisation of a system with its own agency. The rationale is in part that such an agency might provide a form of governance, legal and economic substrate that would be more rational and objective than any human or human-based system might. This disinterested algorithmic substrate would facilitate all manner of rules, currencies, organisations, differences and projects, and in this sense be beyond the political, because it would be able to facilitate any kind of systems design. At other times, such perspectives of

84

New companies are springing up to serve this market of feeding dApps with data, amongst others Oraclize

http://www.oraclize.it. 85

algorithmic mediator take more sinister forms, arguing for why an agency constructed through mathematics and markets represents a necessary and inevitable evolutionary stage and the possible extinction of human beings as a result.86

Here, in particular, I want to draw in the cut of the insensible in order to disentangle and suggest a different ground for debate. When faced with such notions of the necessity of an algorithmic determining agency it is tempting to react by reasserting human responsibility, control and oversight. Yusoff, in her discussion of the insensible and the extinction of forms of life that are beyond sensibility, raises the question of the possibility of a response and responsibility towards that which has not been sensed, has not made itself matter to a specific sensibility. I want to argue that such a perspective might be valid and helpful for understanding not only forms of biological or mineral life beyond sensibilities, but also algorithmic agencies. To further expand on this, I will draw in what is perhaps an unusual take on ideas of AI and autonomous systems, namely animism. I want to suggest that anthropological studies of animism might provide some helpful ways of construing non-human aliveness in techno-scientific fields too. Studies of animism intersect in interesting ways with new materialism debates about how the material world ‘matters’ in ways that exceed assumptions of mute objects for the manipulation of humans. These imply that the ‘aliveness’ and forms of autonomy that are ascribed to such systems do not have to be fully adopted as inevitable (nor even necessarily important) but neither dismissed as speculative fantasy (and thereby assuming full control and agency with humans). Anthropological accounts of animism have since the ‘90s sought to re-describe animist practices in an attempt is to understand how ‘aliveness’ might be understood in specific ways and make sense within particular practices and contexts (Bird-David, 1999; Ingold, 2000, 2011). The emphasis is placed on relations and agency understood as exerting a particular force, mattering by literally making a material difference to a people or person. The specifics matter; this storm, this tree, this rock is related to as a being, because these have exerted some force putting them into relation with a person or a group. There is a relationship of some kind, whether malicious or not, and it therefore matters. This also means that what matters does not rely on some general description (not all storms, trees or rocks matter or are ‘alive’), but is contextual (importantly for networked technologies, contextual here should not assume ‘locality’ necessarily).

Reading the form of autonomy implied in the ambition of Ethereum Smart Contracts and DAOs through these ideas is helpful in several ways. Firstly, it avoids the necessity to counter the idea of autonomous systems by insisting on human control and complete knowledge. Human control and complete overview of algorithmic systems are strenuous claims. Large

86

I met such perspectives from individuals at meet-ups and developers’ conferences in conversation. In some ways it is the logical extension of the idea of ‘mushy humans’, some going so far as to say that if some person was not able to effectively engage with blockchain systems and markets, they were of an inferior genetic strand that should probably die out.

network systems, and the operations of algorithms, indeed comprise elements that are beyond the immediate control, oversight and comprehension (Seaver, 2014; Amoore, 2015; Burrell, 2015). (And that therefore imply already existing practices of working with that which is beyond modelling and certainty). Secondly, the insensible and an animist understanding of relational aliveness also suggests that the affinities and relationships that are formed do not necessarily follow human versus machine narratives that tend to dominate imaginations for how autonomous systems might play out. To close off this chapter, then, I would like to expand a bit more on this second point and explain how recognising such affinities that cut across assumed lines between humans and ‘the rest’, along with the concept of the

insensible, help to counter deterministic reasoning about the most appropriate agency of control, whether human or machinic.

The promise of an algorithmic, neutral determining sensibility that can be known, with mathematical certainty, to have certain properties, is simultaneously a promise of radical uncertainty, a machine necessarily beyond control, out of control. Operating beyond control suggests a particular ‘aliveness’ of an algorithmic determining agency, which then at the same time becomes necessary for the project of establishing a generalised, trustless platform. The two justify each other: a determining agency founded on indisputable proof cannot be controlled by any other agency or it would necessarily be tainted. In the engineering of a deterministic system, there is an awareness of aspects of the system, in particular how it plays out in practice that cannot be fully predicted or controlled. This is also the exact moment at which this aspect of being out of control conjures a certain ‘aliveness’ from this thing that is being built, a force that begins to shape aspects of the world that is not under direct command. This is all fine and well, and importantly, there are corrections; tuning, maintenance and changes that take place in the engineering of systems and protocols such that they begin to align more with the desired intentions and sensibilities. In fact, such tuning, corrections and maintenance reveal much of the desired intentions of such systems in ways that do point towards responsibility. But there is another problem, namely the necessity for the deterministic properties of such systems to be presented as universally mattering. The particular autonomy that is often articulated in Ethereum and Bitcoin is one that assumes universality on the basis of neutral mathematical determinacy. The autonomy and agency of Smart Contracts and DAOs might be considered real enough, in that aspects of these are indeed being built and might very well determine some things and act as a force in some people’s lives. And yet, drawing on a perspective of aliveness from animism, such claims of aliveness do not allow for universal demands that this thing matters. From an animist perspective, and indeed from a radical Baradian interpretation, what matters depends on the sensibility that makes it matter, so to speak. And such sensibilities are not singular and universal, but rather multiple and, as Yusoff points out, also insensible (2013a). Drawing in

ongoing relationship with a field of indeterminate potential. This particular mode of determinacy matters significantly to some, but in no way has to be made to matter for others. The argument I am making then, is primarily a philosophical basis for the particularity of any algorithmic determining agency, as a prerequisite for assessing its particular qualities and assigning a rather limited role for such.