A Theory of Change for ICT4D

Duncan Edwards posted a great piece last week entitled "The revolution will NOT be Open Data" in which he riffs on Gil Scott-Heron's critically-conscious lyrics. Like Duncan I have 'a lingering sense of unease' that all is not right in the world of open development / open data. Like Gil Scott-Heron I don't think we can just 'plug in, turn on and cop out'.
 
Duncan rightly questions the simplistic equation underlying most open development Openness  + ICTs = Development Outcomes.
By way of remedy he echoes Matt Smith's suggests that (open) development needs a more sophisticated theory of change than this – and adds that any such theory of change must factor in power and politics – key elements left ignored in most ICT4D.
 
Just so that we have something to shoot down I will suggest a theory of change in this post, but ultimately I believe what we need to elaborate is a full-blown critical theory of ICT4D that  would enable us to discriminate between PR-puff, development-bluff and people's use of ICT for effecting their own development but maybe I'll leave that for another post....
 
The simplistic Open + Tech = Dev theory of change bears no more critical scrutiny than the claims that "Communication is Development" or of “Info as Aid”. The first striking thing about this equation is that it doesn't mention people. I think it's self-evident that any theory of change for development needs human agency at its heart. For this reason there can be no technology solution to development. After all it is peoples' agency and intent - not their tools - that produce development, is'nt it? As Kentaro Toyama points out, technology - however cunningly designed - can only ever amplify existing human capacity and intent, and can never be a substitute for it. 
Following this logic a starting point for a 'new and improved' theory of change might be:

The application of human agency + knowledge + (open) technology to solve human problems = (open) development 

or

people + knowledge + technology = development
 
However as Duncan suggests, this process does not happen in a vacuum devoid of power and politics. What constitutes knowledge and who determines which problems should be solved are political issues. It is those with political power (and not those with access to data) who determine what interests will be served. Our theory of change needs to redistribute information and power.  
 
Paulo Freire taught us that the 'intended beneficiaries' of development need at the same time to be the agents, authors and architects of that development. This is not currently the case for open data. So-called 'open data' today is only practically open to a tiny technocratic elite who are disproportionately urban, white, middle-class, university-educated and male.
At this point it is the already privileged that are best able to take advantage of open data with the accompanying danger that we may just be adding a new 'data divide' to existing divides? To address this damaging deficit, investment in capacity-building for inclusive participation must move to the centre of 'open' models.
 
Perhaps as a reality check we should only allow that an initiative constitutes 'open development' or 'open data' when its active users programme actors truly represent the diversity of the 'intended beneficiaries' and results in their increased agency?
 

How can we improve on this admitttedly rudimentary theory of change?