The Social Dilemma is a Great Conversation Starter, but What Does the Science Say?

Despite the occasional dash of statistical optimism, it’s difficult to escape the sense that things — broadly considered — are not going well. The pandemic wears on. Protesters fill the streets. The election is imperiled. Wherever one looks, the world (or at least the United States) appears to be on a downward slide. It is tempting, amid this plethora of apparent problems, to seek out a root cause. The Social Dilemma, a new documentary from Netflix, makes the case for a potential culprit: social media, the business models that drive it, and the misaligned incentives that inevitably result.

As an activist call-to-arms, the film is chilling and effective. Using a mix of interviews and dramatizations, director Jeff Orlowski depicts and describes all the headline issues: addiction, distraction, depression, anxiety, social comparison, echo chambers, polarization, radicalization, misinformation, and conspiracy theory. Readers acquainted with the Center for Humane Technology will recognize many familiar faces and arguments. Tristan Harris, the Center’s founder, headlines a long list of technology insiders explaining how they’ve come to view the products they built as destructive.

Harris has stated that one of his great hopes is that, in a world where shared realities have broken down, this film will at least provide “a new shared reality about that breakdown of our shared reality.” But as mixed reactions to the film attest, building consensus about a topic as complex and controversial as this one is difficult in our day and age. If The Social Dilemma works to stir up a larger public conversation about technology and its damaging effects, we hope that the next phase of that conversation can be informed by the best scientific research. In that spirit, below we address some of the issues raised in the film and what the science has to say about them.

Addiction

Dr. Anna Lembke, a Stanford psychiatrist and addiction expert, summarizes her view in the film thus: “Social media is a drug.” Depending on who is saying it and what they say afterwards, this statement can mean many things, from the banal claim that “social media can be fun to use” to the preposterous Independent headline, “Giving your child a smartphone is like giving them a gram of cocaine.” Lembke’s view falls somewhere in between: she doesn’t draw any direct comparisons to cocaine, but she warns that social media plays upon our “biological imperative to connect with other people,” which acts through “the release of dopamine in the reward pathway” -- the same pathway the cocaine acts on. 

Of course, connecting with others in person releases dopamine as well. If social media is an addiction, we have to explain why in terms that go beyond a healthy teenage interest in social life. Psychology of Technology Institute (PTI) advisor Adam Alter is quick to distinguish between substance addictions of the kind fostered by drugs and behavioral addictions like gambling. Social media addiction would fall in the latter category. But as he points out in our interview with him, using the term “addiction” too loosely can lead to the conclusion that a compulsive behavior like breathing is an “addiction” too. Pathology is the crucial factor here: to claim that social media is addictive in the clinical sense, we have to show that it has harmful effects.

In the film, Lembke’s commentary is played alongside a dramatization in which a family attempts to impose a no-screens-at-dinner rule. The attempt fails when the younger daughter in the family smashes the time-locked container the phones are being held in with a hammer. In the process, she breaks her older brother’s phone screen, leading to a deal: if the son can abstain from using his phone for a week, the mother agrees to buy him a new screen. 

Here, the filmmakers make an important point. The powerful algorithms and engagement teams behind social media platforms are personified in the film by a three-person team (all played by Vincent Kartheiser) in a control room. When the son begins his week of abstinence, the team springs into action, sending him a notification that his crush has begun dating another guy. The fact that these dramas of teenage life are now frequently mediated by third parties with independent interests (even if those third parties are faceless algorithms) should give all of us cause for concern. Plausibly, the sight of the notification leads the son to forfeit the deal and check his phone, only to be thrown into a depressive scrolling stupor by the news. 

But the film’s dramatic strategies raise some questions as well. The negative effect on the son’s mood is determined by the bad news in his social life, not by social media alone. Whereas a night of binge-drinking might have negative effects independent of what spurred it, social media platforms are hard to distinguish from the content -- good and bad -- that they host. To avoid a moral panic borne of careless cocaine analogies, we have to be clear about the precise conditions and mechanisms that lead social media to be harmful, when in fact it is.

Wellbeing

Ben, the son in the film, begins a steady descent soon after his night of scrolling. He begins losing sleep and cutting sports practices to stare into his phone. This downward slide is presented as all but inexorable, as the three men in the control room guide Ben’s every choice. The reality is more complicated. A large meta-analysis of social media’s impact on wellbeing by PTI advisor Jeff Hancock revealed little to no effect. Other studies have shown that social media’s negative impact on teens in particular is gender-specific (it is more harmful for girls), and dependent on excessive use. A growing consensus suggests a “goldilocks effect”: both too little use of social media and too much are associated with reduced wellbeing, while moderate use can have beneficial effects.

This shouldn’t be cause for complacency; after all, social media companies do not have any concept of “too much use.” From their point of view, the more we use their services, the better. As Harris points out in the film, we have regulations to prevent energy companies from profiting from our excessive use. To the extent that social media companies have a similar perverse incentive, regulation may be in order. 

But just as “screen-time” is too imprecise a metric to be useful, different types and uses of social media can have very different effects. This shouldn’t be surprising: three hours spent messaging with the love of one’s life over Facebook Messenger likely has more salutary effects on wellbeing than even 15 minutes of indulging in political outrage. Some research has borne this out, but much more will be needed if we want to approach regulation responsibly. 

Echo Chambers

Ben’s woes take a dystopian turn when he gets sucked into the rabbithole of an unspecified conspiracy and becomes radicalized by the “extreme center.” His sister’s efforts to dissuade him fall on deaf ears, and soon his whole world becomes enveloped by propagandists telling him not to trust the system. If The Social Dilemma has a primary thesis, it’s that the information siloing that leads to Ben’s radicalization is unsustainable for democracy. Ben’s case, which has many depressing real-life analogues, is portrayed as an extreme version of what has happened to all of us: recommendation algorithms create echo chambers that erode our shared reality and turn us against one another.

In the popular imagination, echo chambers act as hermetically sealed ideological bubbles that prevent exposure to the other side. On this account, polarization and radicalization occur because we simply don’t know the other side’s positions. In the case of the political left and right, this is not generally true. Most politically engaged people do get exposed to the other side; it’s just that this exposure happens in a context that incentivizes outrage and moral grandstanding rather than mutual understanding. In fact, a 2018 study showed that more exposure to opposing views online actually increases polarization.

As we wrote in a prior newsletter, the philosopher C. Thi Nguyen introduces some helpful distinctions here. Nguyen distinguishes epistemic bubbles from echo chambers. The former function by excluding opposing views; the latter actively discredit opposing views in advance. Nguyen calls this discrediting “evidential pre-emption.” Because opposing evidence is preemptively dismissed as untrustworthy, echo chambers survive and even strengthen in the face of counter-arguments. This makes them exceedingly difficult to combat. Evidential pre-emption can be seen everywhere from cults to partisan cable news.

As portrayed in the film, this process interacts in an especially pernicious way with recommendation algorithms. These algorithms do not have infinite capacities; rather, they excel at identifying particular rabbit holes to which they can then pattern-match our interests. As PTI advisor Stuart Russell points out in our interview with him, one way for a recommendation algorithm to meet its objective of predicting your interests is to make your interests more predictable. And since a political ideologue will be much easier for a recommendation algorithm to satisfy than someone with nuanced views, these algorithms tend to serve us information that makes us more partisan.

Polarization

As Harris explains in an interview on the Making Sense podcast, “we’re about ten years into this mass psychology experiment…[and] actually all of us are kind of running malware and bad code. It’s not just that the other side is wrong; it’s that all of us have been living in such narrow views of reality that we can no longer empathize with each other.” The tricky thing about addressing polarization in a polarized world is that we all tend to see the other side as the problem. Though we can observe the trend of overall polarization over time, this does little to moderate our views. Your echo chamber is my oasis of truth amid the lies.

The Social Dilemma’s “extreme center” conceit is one way of trying to get around this. By evoking the dangers of partisanship without triggering the defenses of either side, the film has the potential to inspire a wide audience to engage online political content more critically. Whatever the inaccuracies of the film’s control room analogy, it is closer to the truth than the folk theory of a neutral feed. We know, for example, that political content that triggers our moral emotions spreads more widely, and that the design of social media platforms incentivize polarizing expression. Just creating common knowledge around this fact may help mitigate some of the harmful impacts of these services.

If we can agree that accelerating outrage and polarization are a problem, we may come to see that we disagree about other things less than we imagined. As a recent paper from Anne E. Wilson et al. suggests, one way social media contributes to polarization is by playing host to the polarized conversation of elites. This in turn creates “misperceptions of division among the electorate, which in turn can contribute to a self-perpetuating cycle fueling animosity (affective polarization) and actual ideological polarization over time.” In other words, conversations on social media can make us think we disagree more than we do, which in turn leads us to disagree more than we have to.

Conclusion

An important part of our mission at the Psychology of Technology Institute is to improve the public conversation about technology and its effects on humanity by transcending the extremes of utopianism and dystopianism. This means listening closely to the warnings presented in films like The Social Dilemma, but also moderating those warnings when the science dictates. We hope that the film will inspire many more people to join the conversation (and dig into the research) about the technologies shaping our minds and our future.

References

Bail, C. A., Argyle, L. P., Brown, T. W., Bumpus, J. P., Chen, H., Hunzaker, M. B. F., Lee, J., Mann, M., Merhout, F., & Volfovsky, A. (2018). Exposure to opposing views on social media can increase political polarization. Proceedings of the National Academy of Sciences of the United States of America, 115(37), 9216–9221.

Crockett, M. J. (2017). Moral outrage in the digital age. Nature Human Behaviour, 1(11), 769–771.

Denworth, L. (2019, May 30). Worry Over Social Media Use and Well-Being May Be Misplaced. Psychology Today. https://www.psychologytoday.com/blog/brain-waves/201905/worry-over-social-media-use-and-well-being-may-be-misplaced

Flaxman, S., Goel, S., & Rao, J. M. (2016). Filter Bubbles, Echo Chambers, and Online News Consumption. Public Opinion Quarterly, 80(S1), 298–320.

Marks, I. (1990). Behavioural (non-chemical) addictions. British Journal of Addiction, 85(11), 1389–1394.

Nguyen, C.T. (2020). ECHO CHAMBERS AND EPISTEMIC BUBBLES. Episteme; Rivista Critica Di Storia Delle Scienze Mediche E Biologiche, 17(2), 141–161.

Orben, A., Dienlin, T., & Przybylski, A. K. (2019). Social media’s enduring effect on adolescent life satisfaction. Proceedings of the National Academy of Sciences of the United States of America, 116(21), 10226–10228.

Orben A. The Sisyphean Cycle of Technology Panics. Perspectives on Psychological Science. 2020;15(5):1143-1157. doi:10.1177/1745691620919372

Przybylski, A. K., & Weinstein, N. (2017). A Large-Scale Test of the Goldilocks Hypothesis. Psychological Science, 28(2), 204–215.

Schroeder, J., Kardas, M., & Epley, N. (2017). The Humanizing Voice: Speech Reveals, and Text Conceals, a More Thoughtful Mind in the Midst of Disagreement. Psychological Science, 28(12), 1745–1762.

Steers, M.-L. N. (2016). “It”s complicated’: Facebook's relationship with the need to belong and depression. Current Opinion in Psychology, 9, 22–26.

Wilson, A. E., Parker, V., & Feinberg, M. (2020). Polarization in the contemporary political and media landscape. Current Opinion in Behavioral Sciences, 34, 223–228.