Shoshana Zuboff talks about ‘Surveillance Capitalism’

Shoshana Zuboff has been attracting a lot of attention (and a few lengthy critiques) for her new book, The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power.

This recorded talk from the Data+Society series of ‘DataBite’ talks provides a quick introduction to her thesis – which is particularly expanded upon in the Q&A section.

Throughout this talk Zuboff makes a persuasive case for arousing populations into action (and resistance) against the rise of the dataveillance economy and the economic logic of continuous data sharing and ‘big data’ computation.

Some of the main take-home points include the following observations:

  • We are living in an age defined by the application of Artificial Intelligence, machine learning and other forms of previously unimaginable computational power to mass datasets. While hugely complex, ultimately these computations boil down to the simple act of making predictions – more specifically, predicting what individuals will do next. This has proven to be extremely lucrative – first allowing for the direction of targeting advertising through to the later ‘personalisation’ of services, and current rise in automated ‘recommender’ and guidance systems.
  • Tellingly, Zuboff argues that these datasets originally stemmed from various forms of ‘collateral’ data. This was system data that was produced through people’s interactions in online environments but that was otherwise not being used. This takes the form of data-logs of clicks, locations and so on – what may once have been referred to in the 1990s and 2000s as ‘digital exhaust’ or ‘digital breadcrumbs’. This is not data that is explicitly generated to indicate the kinds of things that it is nevertheless now being used for.
  • Zuboff contends that the dominant surveillance capitalism that we now have was borne from the early 2000s’ Dot.Com crash, when companies such as Google made the discovery that when this collateral data was subjected to emerging AI and machine learning techniques it resulted in remarkably robust predictions of future user behaviour.
  • There is now a keen imperative in ‘hunting’ this type of data from all our online environments – crucially, only with our implied consent (as Zuboff puts it, asking us for permission would not make for a smart business model). The gold standard for companies is now to be able to produce predictions that can ‘approximate observation’ – i.e. realistic data predications that can allow us to authentically witness the future unfold.
  • Reflecting her roots in a Harvard Business School, Zuboff riffs on the idea of an imminent market in ‘behavioural futures’ where this information is traded and speculated upon – much in the same way that we currently have markets that trade in bonds, stocks and derivatives.


Elsewhere, Zuboff distinguishes between three steps that lie at the heart of the development of this surveillance capitalism logic over the ensuing 20 years, i.e.:

  • Economies of scale (the initial emphasis placed on gathering more data, creating bigger data-sets and forms of mass observation)
  • Economies of scope (the subsequent imperative of gathering diverse forms of information – not just clicks, but people’s emotions, how many steps they take, and so on)
  • Economies of action (the more recent interest in intervening in what people do – nudging people and herding populations toward our commercial outcomes)


Toward the end of the talk, Zuboff also draws attention to the “Radical indifference” of our data infrastructures (what she calls “Big Other” rather than “Big Brother”).  She makes the point that these are systems that do not care what we do … as long as we are producing data!