Online misinformation and propaganda. Are we all involved?

Elections and political votes over the last few years have been plagued with claims of interference and politicians and the media have accused others of spreading lies and “disinformation”. It has often been claimed that social media platforms such as Twitter and Facebook have enabled those with nefarious intentions to sow confusion. But is this any different to the propaganda which has been used by those in power (and those who want power) to encourage people to think and act in particular ways? In a recent paper (official version here (paywalled) and open access version here.) I tried to answer this question.

I think there are some differences and at the heart of this is the role which “normal” Internet users are now taking in producing and/or distributing propaganda type messages.

What has often been overlooked is the different ways in which users are mobilised in order to produce certain political outcomes. While the spread of false messages might seem random and “viral” they are perhaps more directed than they initially seem. This doesn’t mean there is a vast conspiracy with a puppet master at the centre. Rather, political propagandists have developed sophisticated understandings of how social life functions online and have developed methods of subtly “steering” online life to foster certain ideas, biases and actions.

I came across an approach to “psychological operations” (or propaganda) developed by the USSR during the Cold War which they referred to as “reflexive control”. It is a form of “non-linear warfare” which is allows a “controller” to target an opponent with specially prepared information to influence them to make a desired decision seemingly on their own.

This approach is slightly different to traditional forms of propaganda as rather than trying to convey a message to a particular group or individual it tries to affect their decision-making process by manipulating their perception of reality. In order to do this it was required to construct profiles of the targets so that their moral and psychological dispositions could be modelled based on their biography, habits and psychology.

In this scenario a “controller” tries impose their will on a target not by telling them what to think but by influencing their decision-making practices and their perceptions of reality.

“Reflexive control” was developed in the pre-Internet era and was not used for mass persuasion but instead targeted military or political leaders responsible for making strategic decisions. A lot of information on them would be publicly available or obtainable through spying which could feed profiles. These profiles made it possible for spy agencies to play on their interests, biases and insecurities.

But it has become possible to use the “reflexive control” approach on much bigger groups due to the kind of psychographic profiles and targeted advertising which are enabled by social media.

Some of this new kind of propaganda uses directly employed agents to sow confusion and spread “fake news” such as the “trolls” employed by the Russian “Internet Research Agency” who tried to interfere with the EU referendum in the UK and the 2016 US presidential election. Other people are indirectly recruited for such a cause because they find they can make money from it such as the Macedonian teenagers who found that producing pro-Trump and anti-Clinton “click bait” websites with fake stories could make them millions of dollars.

Perhaps most distinctive about the kind of propaganda which happens online is the way in which all users are “unwitting agents of influence” in these campaigns. It is well known that social media sites (such as Facebook) make money by constructing profiles of their users so they can sell targeted adverts to company who want to find customers. These profiles contain the same kinds of information as used by the proponents of “reflexive control” in the USSR during the Cold War. But the data in these profiles has come not from spying but from the users “liking” friends photos of their dogs and sharing news articles.

The former Cambridge Analytica employee Christopher Wiley has said that such profiles were crucial in their political influence operations as they were able to use “off-the-shelf” cultural narratives from the fashion or music industries and connect these with Facebook data to determine which users might be persuadable to pro-Trump messaging. Such targeting is seemingly relatively easy. An analysis by the Washington Post showed that simply knowing which beer someone drinks is a strong indicator of their political persuasion. The ability to “like”, “share”, “retweet” also enable users identified as receptive to such messages to also act as “spreaders” of this content.

As well as allowing us to communicate in new and fun ways allow the “reflexive control” approach to propaganda to be applied not just to prominent and influential individuals but to almost entire populations. This simply wouldn’t be possible without the kinds of data now routinely generated on all social media users. Also, while “normal” citizens have always been implicated in the spreading of messages through word of mouth it is now easier for propagandists to identify particular individuals who are likely to be receptive to their messages and to distribute these is more widely.