A politicotechnical approach

18 Oct 2016

It’s not that uncommon nowadays to criticize the widereaching ambition of the tech sector, what Evgeny Morozov calls “solutionism,” that is, the idea that (often complex social or physical problems) can be solved by the application of recently developed technologies. Put at its most blunt, the complaint would look at something like the Gates Foundation and ask what exactly it is about building a career on buying a knock-off version of CP/M that qualifies Bill Gates in his ambition to change the entire education sector.

This criticism is all the more potent because most IT innovation is crap. This may seem an inflamatory way to put it, but how else can you describe an industry in which only 16% of projects are completed on time and on budget, and where the remaining 50% of projects which weren’t cancelled outright only delivered on average 60% of the planned features (Standish Group)?

One widely studied, because so spectacular, IT project failure is the NHS’s National Programme for Information Technology, which was begun in 2002 and cancelled in 2011, having cost over £10 billion and having pretty much entirely failed to deliver what was supposed to be its central technological innovation, a system for unified patient records across the NHS.

One of the main problems with the NPfIT seems to have been that it was conceived of and practiced as a project to introduce a bunch of technology into the NHS, without much thought being given to what technology the day-to-day practice of the NHS (i.e., treating ill people) actually needed (Clegg & Shepherd). We could probably blame Bill Gates for this, too: the project came out of “a meeting between the Prime Minister and then CEO of Microsoft, Bill Gates, after which the Prime Minister is said to have become ‘hooked’ on the technological possibilities for improvement in the NHS” (Campion-Awwad, Haton, Smith, & Vuaran).

However, we should be wary of taking, as Clegg and Shepherd do, too rosy a view of a view of an alternative, sociotechnical approach, which would consider technology in a social context. Clegg and Shepherd mention one potential advantage to the process adopted by the National Programme for IT:

Because the health service is so big, so complex and so differentiated, it would take forever to get the various end-user communities to lead and agree a project of this kind. As such the IT is used as a Trojan horse to force through some changes in working practices and processes.

What they don’t mention is the kind of changes in working practices and processes that coincided with the National Programme for IT. The changes to the NHS introduced by the New Labour government were specifically neoliberal ones, that is, changes based around an intensification of the internal market mechanisms introduced by the previous Conservative government, organised around competition, measurement of outcomes, and deprofessionalisation of the roles of health care workers.

The point I want to make by bringing this up is that the changes in the organisation of the health service that a properly implemented NHS IT programme could enable would be the result of particular and contestible political choices: they would have been the materialisation of the political victory of neoliberalism over alternative ways of understanding the role of government.

Clegg and Shepherd present their sociotechnical alternative to the techocentric approach in quite fluffy terms, talking about processes that are “owned and led by senior end-users,” (though note they limit this to senior end-users; I’ll have more to say about this in a future post), who introduce “new working practices” as part of a “National Programme for Service Delivery.” But if they simply accept a particular conception of “service delivery” as given, their IT projects will contineu to be a “trojan horse” for political projects they don’t consider.

It seems to me that the study of information systems needs to pay attention to an insight from critical theory: that the social is always organised according to power relations, and so any sociotechnical approach is also a politicotechnical approach. It might be, as in the sociotechnical approach outlined by Clegg and Shepherd, a politicotechnical approach that is simply silent about, and so accepts, existing political relations. But I think there is, or should be, space for what we might call a critical information systems, that attempts to be self-aware about the political choices that are embodied in the information systems we are building.