The Social Machines Mindset Emerges
For years computer technologists have been trained to focus on the design and engineering of computer systems – to build systems which address “user requirements”, where those users are somehow external to the system. Thinking has changed a bit over the years and sometimes now (not often enough still!) those “users” are much more closely involved in the design, for example in the Web 2.0 approach – where perpetual betas can be achieved with websites in a far more agile fashion than traditional models of software distribution and installation.
Now we are seeing a new approach. Instead of thinking of the computers as independent systems used by users, we are thinking of the socio-technical system of computers and the many human beings engaged with them – the Social Machine. Take Wikipedia for example, with its contributors, editors, and the ethics and etiquette that have become established as it evolves. The “programming” of this overall system has been socially constituted – it didn’t come out of requirements capture, traditional software engineering and development. Similarly the social machinery around twitter hashtags was enabled by the design but has evolved through creative use by “the crowd” (and, according to the BBC at least, has changed the world).
The Galaxy Zoo citizen science project is another powerful example of successful crowd-machine symbiosis (and its co-founder Chris Lintott has just been honoured as an internet pioneer). The computer system is using humans for “data reduction” because their cognitive apparatus is beyond the state-of-the-art image processing (at the moment). Is it just humans as subroutines, enslaved by smart task allocation and training algorithms to support distant scientists? No, the humans are using Galaxy Zoo to make new scientific discoveries – not just the scientists, but the citizens who are engaging in the science, interacting through discussion fora. The behaviour of this social machine has been emergent, and both are empowered. And this Social Machine is so successful that we now have the Zooniverse project launching new citizen science projects routinely: thus emerges the social machine which is a factory for social machines. Zooniverse, and citizen science projects more broadly, are receiving study as important and fascinating examples of large-scale socio-technical systems.
So the Social Machine is not just a computer system with a bit more “social” thrown in; it is fundamentally a system designed and realised with due attention to both the humans and the computers*, its behaviour ultimately defined by the ingenuity of the crowd. Anecdotally a new generation of web and app designers are already heading this way, so perhaps we are simply giving a name to something already out there - as Web 2.0 once did. But I fear it will take a while for computer science, and indeed for software engineering, to reimagine itself and embrace (codify?) this emerging practice (where the word "user" is rightly deprecated). Social Science, meanwhile, has a history and indeed extensive literature on socio-technical systems: for social scientists, these Social Machines are distinctive new categories of interest.
Observing successful social machines is important to help us understand how they came about, their lifecycle and what we can learn from that (like the decline of wikipedia?); observing failed ones is just as important. But the challenge for all who understand the world as an ecosystem of interacting social machines is to figure out how we design a new social machine and release it into that ecosystem, as a purposeful intervention which evolves to achieve the desired outcomes. This mindset sees the world as a Social Machines laboratory and the study of Social Machines (in projects like SOCIAM and Smart Society) is not just about observation, it is about design and construction, about coproduction and communities solving problems, and about power, trust and responsibility. Methodologically we are working in the wild - the social machines jungle even - conducting mode 2 Web Science which would perhaps be more familiar in an iSchool or business school than most CS departments.
So our exciting future is in an overlapping space which seems uncomfortable for computer scientists (the problems don’t yield to traditional instincts for solutions), is a new emphasis for social scientists (addressing purposeful design and construction at crowd-scale) and which potentially engages in any and every discipline touching on the socio or the technical. The Social Machines mindset is already becoming established today as part of Web Science, sitting quite comfortably in our Digital Economy world too, and I hope that others will engage and embrace (and submit papers to the Web Science conference!)
Today it's great to see Digital Social Research modules in Social Science courses, but now we need to see Social Machines modules in Computer Science and elsewhere. If I'm right about the mindset then how long till we see Social Machines professionals? And we might also ask whether Social Machines expresses a new discipline which supersedes our historical silos.
* That humans and machines should receive symmetrical attention does not, however, say that humans and machines are equal. Rather, the intention has been that machines should do what machines are good at to enable humans to do what humans are good at; i.e. this is about empowerment of humans and especially human communities. The original definition of Social Machines by Tim Berners-Lee in his 1999 “Weaving the Web” book says “Computers can help if we use them to create abstract social machines on the Web: processes in which the people do the creative work and the machine does the administration…”; Hendler and Berners-Lee also coined the phrase "humanity in the loop". While this is an important historical and perhaps cautionary note, the distinction is increasingly blurred: as automation increases (and the footprints of humans are massively captured) we increasingly see “bot or not” confusion between human and machine. Indeed the reCAPTCHA social machine converts the need for bot detection into energy to transcribe digitised text.