Last week, Oxford’s School of Geography and the Environment’s Political Worlds research cluster (thanks for funding the event!) hosted an event I and several others organised around ‘data’s dirty tricks’. As chair, I had no idea what each panellist was going to speak on, which made it both a challenge and equally thought provoking – more information here on a previous blog post. Below is a summary as a chair, and does not necessarily reflect a truthful account of the event, and any errors in the below are mine.
We had three panellists: James Ball (Journalist and author of Post-Truth: How Bullshit Conquered the World), Dr Lina Dencik (Director of the Data Justice Lab at Cardiff University and co-author of Digital Citizenship in a Datafied Society), and Dr Vidya Narayanan (Director of Research at the Oxford Internet Institute’s Computational Propaganda Project). All contributed a different dimension – broadly from the media, from those tracking propaganda and issues of surveillance and privacy. Some tweets from the event:
— Meera Selva (@MeeraSelva1) November 15, 2018
— Meera Selva (@MeeraSelva1) November 15, 2018
What came out most for me is, what is value? This was a common theme; whether from James Ball’s understanding of fake news and how the media attempts to discern what is valuable or not (i.e. more likely to be truthful, deceptive or indeed outright lies). From Vidya Narayanan, it was clear a value emerged in talking about whether Russian bots, for example, in the ‘Brexit’ process had been able to have a decisive influence (it seems they did not contra popular perception, they only amplified – or in her words lead to further polarisation of – existing views of groups). And from Lina Dencik a more fundamental critique of the value of global monopolies such as Facebook and whether they indeed provide a value at all (her discussion of how these companies are steering the debate, and are thus shutting down political debate of their actions was exceptional. Why do we have to respect their value frameworks at all?)? All three provided an excellent overview of different aspects of data’s dirty tricks – and how, I think, data has become one of values. In this sense, how is the collection, processing, and decisions made upon data all laden with different forms of value – and how do these interact and become conflicted in different spaces? The spaces of Silicon Valley are different to Westminster, and even these are different to the values of how those who are on Universal Credit see their data being used, tracked and analysed.
The core ‘surprise’ was that data’s dirty tricks are actually quite tricky. Cambridge Analytica, the firm that took lots of personal data from Facebook and used in political campaigning, was actually pretty poor at addressing the issue of changing peoples’ minds and voting. No more so than on Brexit, with perhaps a better influence on the 2016 US Presidential election. Trying to convert selected data and derive particular forms of value are hard – whether you wish someone to buy a product or vote a certain way in an election. No doubt, there are perhaps some avenues that data have been used adversely – but as Ball pointed out, it was the hacking of the US Democratic National Convention (DNC) by the Russian state hacking group CozyBear (APT29) that released emails relating to Hillary Clinton, that more likely swung the election. This is not to say that computing and hacking cannot be influential, but that data’s dirty tricks may not what they’re all cracked up to be. This is reinforced by Narayanan’s work on Russian bots which showed they are semi-automated and rather poor at directing people in certain ways – only polarising those in different groups away from one another – but maybe that’s enough, to cause polarisation?
Whether we have ‘strong’ organisations also cropped up, with Dencik arguing that due to austerity, there had been a weakening of the state to counteract the demands of tech companies. This leads governments and other organisations to accept their demands, citing NHS contracts with Alphabet’s Deep Mind that took data with little to no patient consent. Therefore it is not only about individual consent about data but thinking about the collective privacy issues that emerge when data is used in these ways. Yet, Ball was keen to emphasise that the mainstream media is actually the main proponent of fake news, not social media, and that it is up to them to do more fact checking but the stresses of journalism make this hard. One thing he said was that we should be proud that the UK has the BBC – as this provides a pivotal place in which to challenge inaccuracies and restrict filter bubbles… However what is distinctive about data is that it has an ability to move in ways previously impossible – and that is the new challenge, one of speed and distribution, rather than one of distinct difference.
I think the discussion left us at two avenues; one where the contortions of social media data do very little, but that the political economies (riffing off Dencik) of the use of data are challenging conventional political decision making. What I find interesting is the recent focus on the former (Facebook, elections, and so on), but little on the everyday issues of Universal Credit, NHS contracts, and outsourcing that are based on the sharing of data without appropriate consent. Hence, the dominant focus on data’s dirty tricks should perhaps switch to focus on the latter – on asking what are the politics behind the use of data rather than on how data can influence elections (as it turns out they do very little). Data’s dirty tricks, to me, seem to be laden with power as much as they ever have been.