What do networks of humans and machines actually do?
We expend a lot of time and energy, especially in a project like HUMANE, trying to understand the ‘what’ and the ‘how’ of human-machine networks, but it is during a workshop such as the recent, excellent, discussions in Oxford that bring to the fore the question of ‘why’.
We are apt to think of the machines in the network as the important feature – after all, the humans have been there all the time, it is the machines that are the innovation. Aren’t they? Maybe not. As Eric Meyer reminded us at the start of his talk, people have been building machines ever since they climbed out of the trees and started banging rocks together. We may not think of a piece of bent stick as a machine, but the use of a tool to dig furrows and plant seeds heralded a major social shift from nomadic to agricultural lifestyles.
Dave De Roure furnished us with more examples, citing the printing press and its social impact in 15th Century Europe leading to the libraries and social records we have today. So does this make the printing press a social machine, in line with the definitions coming from Tim Berners-Lee et al., where he defines social machines as abstract entities living on the web that do the ‘heavy lifting’ of administration, leaving the people to be creative? Or is the concept more abstract still? Whereas the plough-share enabled the people using it to be more productive by making a task more manageable, it also permitted a social change as a consequence of the introduction of a different way of life that was not possible before the machine arrived.
Similarly, but perhaps not so obviously, the printing press caused social change. People could write and distribute their ideas before the printing press arrived, but if they wanted to distribute their ideas widely they were reliant on monastic scribes to create copies. With the arrival of the printing press it became very easy to replicate and distribute ideas in print without involving the monks. This is very much in line with the new forms of social process that Berners-Lee also associates with social machines, and has obvious corollary with the social changes brought about by the rapid expansion of social media at the start of this century.
Of course, we have to ask whether all such changes are beneficial, and who defines what ‘beneficial’ is. Each new technology-led innovation ushers in a Utopian ideal in which the human beneficiaries are enabled to achieve idealised goals – or at least that is what the technologists behind the innovation would have them believe. What we see in reality, whilst not necessarily dystopian, is nonetheless very far from this idealised world. There is, and always will be, a huge difference between the way the humans behave and the way that machines behave. No matter how complex the machine, and how closely it appears to mimic human thought, a machine will never be human, it will always be a machine.
The dystopic view of our developing relationship with machines comes not from machines developing some kind of emergent consciousness and taking over the world, but from the behaviour of the people who exploit them or rely on them. Machines are a product of their design and programming – they have limitations. People, on the other hand, are driven by their very nature to explore outside the boundaries of experience. They don’t ask ‘what does this machine do’, they ask ‘what can I do with this machine’.
Vegard Engen introduced the concept of ‘intentionality’ as a distinction between the ‘agency’ exhibited by machines and the ‘agency’ exhibited by the humans in a network. Humans will intentionally set out to get the machine to do what they want it to do, whereas the machine will only do those things that are within its design parameters.
In the descriptive model presented by Brian Pickering ‘Human Behaviour’ takes centre stage, usurping the earlier focus of such models on the technical capability within networks. This is an important shift of emphasis taking place in the study and understanding of human machine networks, including as it does the social science and humanities component as an intrinsic part of network functionality.
In her review of the roadmaps being developed by the HUMANE project, Eva Jaho talked about policy and regulation as well as technological development – reflecting the need to manage the behaviour and activity of the people in a network whilst recognising that evolving technology allows for emergent beneficial behaviour that could be supressed by over-enthusiastic regulators. We should remember that machines operate on the principle of prescription – they do what they are designed to do – whilst people operate on the principle of proscription – they will do anything they can get away with unless they are prevented from doing it.
Dave DeRoure reminded us that people are subversive – they will be inventive to get the machines to do what they want to do, not what the designers expected the machines to do. The best networks are the ones that celebrate and encourage the inventive ability of humans – Grant Miller provided the example of Zooniverse and its ability to satisfy the higher human ideals of curiosity, satisfaction and achievement whilst eschewing any financial reward.
So, I will return to my original question of what human machine networks, or social machines, actually do. Gina Neff talked about symbiotic agency, reflecting the developing understanding of networks coming out of HUMANE.
Humans and machines work together to achieve a human-defined goal. Different humans within the network may have different goals, leading to conflicts and battles such as those described by Taha Yasseri in his studies of Wikipedia, but this is a result of human nature, not machine intervention. Human machine networks and social machines allow people to do what people do best – communicate, explore, discover, invent, manipulate, subvert and revolutionise.
People have a symbiotic relationship with the machines they invent – but they always have done. Where machines come to dominate or control lives it is only because we have allowed them to do so. We lay ourselves open to Perrow’s ‘Normal Accidents’ but, as Perrow describes, they do not arise because of the technology but because of human reliance and organisational failure. Our understanding and appreciation of the value and benefits of human machine networks must be based on their social context and on the resultant behaviour of the people forming part of the network, we can no longer study networks as purely technological artefacts.
There is no other ghost in the machine than the people who live within it, who seek to achieve their goals and ambitions, their wants and needs in symbiosis with machine capabilities. And this is what human machine networks do – they give us the power to be more human and to do better what we, as humans, have always strived to achieve.