To what extent is the aim of human society to maintain its equilibrium, as the POSIWID principle would suggest? There is a line of French thinkers who resist the universalism to which some schools of cybernetics aspire, and see the construction of social norms as political rather than teleological or quasi-biological. Xavier Guchet traces the position of Canguilhem and Simondon back to Bergson’s The Two Sources of Morality and Religion
where it is argued thatGuchet 2012closedmorality and religion are without a doubt morality and religion of conservation. Their function is to preserve the stability of the existing social order. On the contrary,openmoralities and religions have the function of returning human societies to the élan of creation, of unmaking the existing social order, inventing another order and becoming something else.
Whereas for Foucault and others, social norms are constructed to protect society from pathological variations that might threaten it, Simondon focused on invention and the creation of new norms.
(I am more familiar with Popper's notion of Open Society than with Bergson's. There appears to be some difference between the two notions, but I haven't done enough reading at this point to be able to explain the difference.)
One way of talking about these questions is in terms of programming. Simon Mills quotes from a book by James Beniger, distinguishing between control (purposive influence toward a predetermined goal) and programming (setting of the goal to be achieved). Mills carries out a close reading of one technology advocate (Sandy Pentland), showing that the success stories of big data are largely based on relatively closed or autopoietic systems, delivering some degree of technocratic efficiency and resilience, but failing to answer the more fundamental question - what is the purpose of society as a whole. Where do the goals come from?
Another way of talking about these questions is in terms of organizational learning. Chris Argyris introduced the distinction between single-loop learning and double-loop learning, which very roughly corresponds to Beniger's distinction between control and programming. While single-loop learning uses simple feedback to improve the performance of a system relative to a fixed goal, double-loop learning allows for the modification of goals in the light of experience. Advanced technologies such as machine learning are not limited to single-loop learning, and may be able to do some limited double-loop learning, in suitably controlled environments. (To go beyond this, we may need some notion of triple-loop learning. But see article by Tosey Visser and Saunders problematizing such labels.)
Sometimes, technologies and sociotechnical innovations are spoken of as ethically and politically neutral instruments, which can simply be used to maintain established socioeconomic and cultural goals. So that falls naturally into the "closed" model of society identified by Bergson. But if technologies and innovations (sometimes the same ones) are described as disruptive, this seems to imply a more "open" model of society.
Obviously there are ethical issues both ways - whom does the disruption serve, but also whom does the preservation of the status quo serve?
Henri Bergson, The Two Sources of Morality and Religion (1932)
Geoffrey Bowker, How to be universal: some cybernetic strategies, 1943-1970 (Social Studies of Science 23, 1993) pp 107-127
Philip Boxer, Triple-Loop Learning (Asymmetric Leadership, 8 January 2007)
Xavier Guchet, Technology, Sociology, Humanism: Simondon and the Problem of the Human Sciences (SubStance #129, Vol. 41, no. 3, 2012)
Simon Mills, Simondon and Big Data (Platform Journal of Media and Communication, Vol 6, 2015) 59-72.
Alex "Sandy" Pentland, Reinventing Society in the Wake of Big Data (30 August 2012). Professor Pentland is also mentioned in John Thornhill, Trustworthy data will transform the world (FT, 5 March 2018, paywall)
Karl Popper, The Open Society and its Enemies (Routledge 1945)
Paul Tosey, Max Visser and Mark NK Saunders, The origins and conceptualizations of
triple-loop learning: A critical review (Management Learning 2012 43: 291 originally published online 2 December 2011)