Since it was released in early September, the Netflix documentary The Social Dilemma has attracted much attention (Netflix/Exposure Labs: Dir: Jeff Orlowski 2020).
It was for a time the top streaming film on Netflix. It argues that social media such as Facebook and Instagram, along with Google and other digital platforms, have turned smartphones into attention seeking devices, leading to mental illness, undermining democracy, and eroding the truth through disseminating fake news. The Social Dilemma makes this case strongly since those who argue the case are former senior social media staff, many of whom shaped these platforms.
The film argues that social networks are insidious engines of behaviour modification. To pursue their business model, the social media platforms need to be attention machines maximising the time users spend on them. This contrasts with the earliest internet social media of the early 1980s - Usenet newsgroups, largely un-moderated bulletin boards of comments on a given topic. These were not always healthy places - flaming, trolling, spam and sockpuppet first became commonplace terms around newsgroups. But ultimately, the degree to which an individual used these groups was determined by their enjoyment of the discussion. If these forums turned toxic a user would simply walk away.
Modern social media is unlike this, not simply through their ubiquity but because they have used psychological techniques to modify people’s behaviour. Notably they use intermittent positive reinforcement, the same basis on which slot machines relieve people of their money. Straightforward positive reinforcement describes the process of a subject being rewarded after a certain kind of behaviour thus making the subject likely to repeat that behaviour. But this would not work with a slot machine – if you won each play it would become dull and the behaviour would not be reinforced. Intermittent reinforcement gets around this by not rewarding every instance of the desired behaviour, thus creating lows (associated with losing/no reward) to make the highs (winning/reward) greater. Social media platforms consciously use this technique through intermittent likes, shares, notifications and other rewards of social recognition, acceptance, feeling wanted and liked.
Some call the resulting pattern of behaviour “internet addiction”, although the usefulness of such a term is debatable. More importantly, intermittent rewards lead to stress and anxiety when behaviour is not rewarded which can (at least) exacerbate mental illness. This is the view of one of those who introduced the “like” button to Facebook, Justin Rosenstein, as he attests in the film. Research from the Comprehensive Clinical Trials Unit at UCL has shown an increase in rates of clinically identified anxiety, particularly in 18-24 year olds, with rates in the UK trebling in the 10 years from 2008 (the year after the iPhone was introduced) although rates had been stable for the ten years prior to that. The UCL researchers suggest that this may be caused by growing inequality, austerity, Brexit or climate change. However, similar trends are found in the USA and elsewhere and there is a clear explanation as to why social media might have this effect.
The attention machine is not an end in itself but creates the audience and data allowing the platforms to make money. As many of those interviewed in The Social Dilemma state, this is far from traditional advertising. Traditional advertisers know that an advert placed during, for example, Love Island will be seen by an audience with a particular demographic profile; it will nonetheless be seen by a diverse group. Social media adverts can be directed at very specific groups (say those who have shared anti-vaxxer posts). Moreover, the platforms gather a raft of behavioural data on their users to create models of how users act (user with characteristic X behaves in way Y) including whether they react to particular adverts. Machine learning and the resulting algorithms mean that advertising is honed to a degree that it is more akin to behavioural modification (creating a precise environment to encourage clicking through on an advert) than traditional advertising.
This business model is not simply a layer placed on top of an unadulterated social media feed, it determines that feed. Modern social media are not 1980s newsgroups that simply present a list of posts on a bulletin board. They mix feed from posts by users’ “friends” with other digital content (Facebook has lessened the amount of news in their feed recently but many people still receive much of their “news” from social media). What a user is presented with is driven by the corporate needs of the social media company to maximise attention and condition the user’s behaviour in line with the advertiser’s requirements.
One of the talking tech- heads who drive the film’s analysis, Jaron Lanier, makes a good point about the impact of this by comparison with Wikipedia. When someone looks at Wikipedia they see the same pages as everyone else, material that Wikipedia moderates for accuracy. Wikipedia is crowd funded and carries no adverts, so its users are its customers. Now imagine that Wikipedia was funded by advertisers. In order to maximise revenue Wikipedia might present different versions of a page to different users, one that (based on a mass of data on that and other users and their past behaviour) nudges them towards the outcome the advertiser was paying for. What is seen on this commercialised Wikipedia page would not be shaped by Wikipedia’s belief in accuracy but would be driven by third party funding. You do not have to imagine this - it is called Facebook.
It is worth remembering that Wikipedia content is not truly “user generated” – most of the material is written by people who are experts in their field, often academics. Unlike vloggers and social media influencers, they receive no reward or recognition. It thus relies on an infrastructure of content generation largely exogenous to the digital world, that of experts whose only reward is (at least in some approximate way) their belief in producing accurate material. Wikipedia is far more an exemplar of creating an intellectual commons than anything corporate social media can aspire to.
The Social Dilemma is not without its flaws. The documentaries exposition is interlaced with a dramatised fictional family of social media users. This attempts to make its message understandable to the less tech-savvy elements of its audience but elements of this, particularly the personification of social media as three little men with computer screens manipulating an avatar of “their” user gives social media’s algorithms agency, personal focus and intentionality which (as the talking tech-heads elsewhere in the film clearly state) they lack. Social media is not intentionally immoral (although some who exploit it are) but rather it is carelessly amoral.
A further important issue is that the film does not place social media into any broader social context. For example, it holds social media responsible for the rise of right-wing populism. While such populism undoubtedly uses social media, it has its roots in a reactionary response to neoliberal globalisation from the 1990s onwards, aided by the historic defeats of the working class and boosted by the financial crisis of 2007-8. All of this pre-dates the rise of social media. The link between populism and social media is anyhow tenuous in some cases – in Britain the demographics of UKIP/Brexit voting are the mirror image of social media use (UKIP voters are older and tend to be non-metropolitan, less educated and more male: social media users much younger, metropolitan, more educated and more female). There are other examples of the rise of social media being coincident with changes but not the cause of them: Danah Boyd writes in It’s complicated (2014) of young people in the USA turning to social networking as public spaces became increasingly inaccessible and parents unwilling to allow younger teens freedom to go out.
Finally the film asks what should be done about this. Ultimately a socialist answer would be based on collectively owned services under the democratic control of its workers and, more particularly, its users (but certainly not state run social media proposed by Jeremy Corbyn when leader of the Labour Party). Reflecting the free-market libertarianism of the tech sector, even the critical wing of it represented here, The Social Dilemma suggests no move in this direction nor any form of radically democratic answer. Even Lanier (an older generation of geodesic dome building pan-pipe playing tech hippy) while calling for people to delete their social media accounts sees this only as pressure on social media companies to change. Others in the film fall back on market reform through state regulation. While state regulation is particularly problematic – states are among the worst abusers of the opportunities for manipulation offered by social media – some form of state regulation to curb the most egregious aspects of social media is the only near-term proxy (however imperfect) for democratic control of social media. Nonetheless, The Social Dilemma remains a sharp critique of the potential of corporate digital media to do harm.