Social media itself is neither inherently good nor bad for democracy, but it can be either depending on who is using it and for what ends. When social media first burst onto the political scene about a decade ago, it was hailed as “liberation technology” that would spread democracy around the globe. People who did not have access to mainstream media could coordinate and collaborate with one another outside the control of an authoritarian state.

Of course authoritarian regimes eventually woke up to the threat posed by pro-democracy online opposition, and developed a suite of tools to confront this, including offline responses (e.g., imprisoning bloggers, changing liability laws), restricting access to online content (e.g., the great firewall of China), or trying to influence the online conversation. This last response option led to the development of political bots and trolls.

Ironically, the very same affordances that made social media useful for pro-democracy activists also make it a valuable tool for those who harbor anti-democratic sentiments in democratic societies. Prior to social media, if you were the only one in your county who supported extremist views regarding the overthrow of the United States government, organizing with other likeminded but geographically dispersed compatriots would be a costly activity. Social media drastically reduces these costs and allows such individuals to more easily find each other to organize and collaborate. Moreover, the tools developed by authoritarian regimes to influence their own online conversations — online trolls and bots — can also be used by small numbers of extremists in democratic societies to amplify their presence online, making their positions appear to be more popular than they might be.

Another liability of social media for democratic societies is that it exacerbates a trend that was already developing on the Internet before the arrival of social media, which is the removal of traditional gatekeepers from the production and distribution of “news.” Prior to the Internet, news was in the domain of professional journalists and there were powerful gatekeepers in the form of editors and publishers. While this may have also prevented more progressive messages from entering mainstream media, it undoubtedly also blocked extreme anti-democratic voices as well, in addition to enforcing a certain level of quality in news reporting. The Internet lowered the barrier to publishing news dramatically, but social media accelerated this process. Social media exacerbated the premium placed on news that delivered “clicks,” highlighting the appeal of certain types of news — including blatantly false news — with particular features, including appealing to partisan identity.

When we turn to more specific accusations, however, such as whether social media is exacerbating political polarization, the picture is a lot less clear. There are theoretical reasons to think that social media could exacerbate political polarization by placing people in echo chambers where they are surrounded by likeminded fellow partisans who denigrate those on the opposite side of the political spectrum. But the evidence for the existence of echo chambers is much more mixed than most people assume, and there are reasons to think that social media actually exposes us to more people with diverse political views than we encounter offline.

There are also reasons to think that social media may privilege the type of news that leads to people becoming more politically polarized, and a recent study demonstrated that taking people off of social media for a month did show lower levels of dislike for out-partisans in the U.S. However, in new, under review, research from the NYU Center for Social Media and Politics, we find that taking people off of social media at a politically fraught time in Bosnia-Herzegovina actually increases ethnic out-group hostility, and that this effect is largely driven by people who live in ethnically homogenous areas of the country. This suggests that for some people, their online connections may actually expose them to a greater diversity of views than they encounter offline in the course of their daily life.

If we look over time, it is clear that political polarization has increased both in the U.S. and globally over the past decade, a time period that coincides with the rise of social media. But there are a lot of other things that have occurred in that time period, including globalization, economic stagnation in the aftermath of the Great Recession, the Syrian crisis and resulting refugee flows in Europe, rising polarization in the United States, a renewed awakening around issues of social and racial justice, and, most recently, a global pandemic. We need to be careful about assuming correlation equals causation in the case of social media — it may very well be the case that political polarization, particularly in the United States, would have been growing in response to long-term economic and demographic developments even absent the dramatic technological changes that have taken place at the same time.