We’re in a kind of “hair on fire” moment with respect to the public mood about social media’s effect on democracy. Popular political commentary reflects this pervading negative mood with headlines such as: Do Social Media Threaten Democracy? Can Democracy Survive the Internet? How Twitter Killed the First Amendment. Your Filter Bubble is Destroying Democracy. 

The growing perception is that “the internet” generally, social media specifically, and dominant American platforms like Facebook and Twitter, even more particularly, are eroding the quality of discourse necessary to sustain democracy.

Citizens and governments are justifiably alarmed over the perceived deleterious effects of social media across multiple dimensions, ranging from creation of echo-chambers that increase polarized political discourse; to serving as megaphones for populist demagogues, terrorists, and authoritarians; to eroding the business model of traditional journalism, which is supposed to play an important watchdog role in democracy. Perhaps the most terrifying threat comes from the “weaponization of information”—manipulative use of social media by foreign governments reaching across borders to “hack discourse” and affect the outcome of elections.

This is a dramatic and dire turn of events: We used to assume social media and the general trend toward democratization of the means of distributing content were net positive forces for democracy: that social media disproportionately empowered civil society, expanded citizen access to information, and enhanced civic engagement. Now, we fear that social media is turning free speech against free society and actually undermining the integrity of democratic elections. The sad fact is that malign actors have figured out how to use social media very effectively against democracy, both inside and outside the United States.

But crafting effective interventions is much more difficult than expressing fear and outrage. The most basic point for democratic governments to keep in mind as they develop solutions is to be careful not to undermine democratic values in the name of protecting democracy.

The political reality is that policy-makers are facing pressure to “DO SOMETHING” about the negative effects of big digital platforms on democracy, without fully understanding all the dynamics at play. Policy-makers don’t have a clear conceptual framing for the various negative effects of social media they seek to reign in, much less a reliable empirical base for assessing those effects. The risk is that well-intentioned governing actors will move to “protect democracy” without thinking through the unintended negative externalities of their proposed solutions.

The core challenge we face is to quickly find effective ways to combat real threats, without undermining our own democratic values like freedom of expression and access to information.

As a starting place, before moving down a path to crafting solutions, policy-makers need to differentiate between platforms and functions within the same platform, rather than lump all “digital platforms” into one category. “The internet” is not the same as “social media.” Social media is not equivalent to search engines. Twitter bots are not the same as Facebook personalization algorithms. User-generated content simply posted on Facebook should not be conflated with newsfeed content pushed to the top of the feed by algorithms.

Second, the various ills in the digital information ecosystem and their deleterious effects on democracy should not all be collapsed into one category. Much of the animating energy driving policy makers to address the “problem of social media” flows from national security concerns about foreign information operations intended to disrupt democratic election outcomes. Foreign disinformation campaigns are not the same as the creation of political filter bubbles or the disruptive effect of social media on the business model of professional journalists in democracy. These distinctly different policy concerns will need different policy solutions.

Furthermore, much of the commentary seems to merge effects driven by inherent features of the technology with effects caused by malign uses of the technology by bad actors who manipulate platform tools for nefarious purposes.

For example, a core critique of Facebook has been that personalization algorithms and micro-targeting capacities shape access to information, create echo chambers and reinforce tribal mentalities, all of which intensify polarization and animosity in society.  (Apparently, the underlying empirical evidence with respect to Facebook’s polarizing effects is still being sorted out by independent researchers.)  BUT, even assuming the polarization effect of Facebook is substantial, the echo-chamber problem is distinctly different from the intentional manipulation of social media by foreign actors seeking to affect election outcomes. Delineating between inherent features and malign uses will help us define platform responsibilities and yield better solutions.

This is a pivotal moment for democratic governance in the digital realm.

To say that digital technology has disrupted society is a cliché. Yet, we are only just starting to grasp the radical break we are facing with respect to our legal institutions and norms, now that digital tools pervade our social relations and information flows. Policymakers are struggling to deal with this transformed information ecosystem, where existing doctrine and regulatory approaches don’t work. The speed, scale, and extraterritorial reach of “bad” online speech makes it qualitatively different from dangerous speech in the non-digitized realm, so we need new tools and rules to counter it. In American parlance, to say that we need “more speech to counter bad speech” just won’t cut it.

But there is an additional wrinkle on this already complex challenge: characterizing the nature of speech in the digital realm and whether it can be legitimately restricted in democracy is itself a very slippery subject matter. We are dealing with a kind of shape-shifting phenomenon where hate speech can look like core political speech that must be protected in democracy, when in fact that speech might be part of a foreign disinformation campaign intended to polarize American political discourse.

We see a bizarre interplay between domestic fringe hate groups and foreign anti-democratic actors, troll farms, and bot armies—mixing political themes with hate speech to sow seeds of discontent, mistrust, and polarization across borders. The hard part is that polarizing speech sometimes looks like it should be protected political speech in a democracy when it’s actually part of a sophisticated information operation we need to protect against.  This has created category confusion for us all.

Here’s the dilemma for democratic governments and policymakers:

Protection of free expression is now coming into tension with protection of democratic processes.  Protection against the pernicious effects of free speech and access to information have become top national security and cybersecurity concerns.

While we sort through this category confusion and develop new doctrine for a dramatically different information ecosystem, we must NOT give up on core democratic values. Our challenge is to stop malign actors while protecting free expression. To do this we need to internalize a “do no harm” mentality and think through the negative externalities of any interventions, especially when those policies might be replicated and go global.

During this period of tumult, democratic governments must not fall into the lazy and cynical view that they can no longer afford to protect free expression because of the dangers of social media. If they do, they will become unwitting participants in the erosion of democracy and help finish the work that foreign information operations started.