Four recommendations about what social media can still do to protect elections
Kreiss, D. & Yael, E. (2020). Additional steps platforms can take to protect the vote. Center for Information, Technology, & Public Life (CITAP), University of North Carolina at Chapel Hill. https://citap.unc.edu/additional-steps-platforms-can-take-to-protect-the-vote/
On July 30th, a small working group of experts on misinformation, platforms, and voting met to analyze platform policies and enforcement of electoral misinformation. The goal was to provide realistic and effective recommendations for additional steps the platforms can take now to combat misinformation about voting. That very morning, President Trump tweeted misinformation: “With Universal Mail-In Voting (not Absentee Voting, which is good), 2020 will be the most INACCURATE & FRAUDULENT Election in history. It will be a great embarrassment to the USA. Delay the Election until people can properly, securely and safely vote???” Neither Twitter nor Facebook took action on this post. A Facebook spokesperson said they did not take the post down because they do not fact check politicians (directly contradicting earlier assurances from Mark Zuckerberg that their voter suppression policies apply to everyone), and Twitter claimed it was a “broad, non-specific statement” in accordance with their policies.
With Universal Mail-In Voting (not Absentee Voting, which is good), 2020 will be the most INACCURATE & FRAUDULENT Election in history. It will be a great embarrassment to the USA. Delay the Election until people can properly, securely and safely vote???
— Donald J. Trump (@realDonaldTrump) July 30, 2020
This example demonstrates the complexities of content that does not fit clearly into defined electoral misinformation and voter suppression policies but that sows distrust in the electoral process. With the election just months away, there has been a surge of false claims on Facebook and Twitter about the process of voting, the security of mail-in ballots, and the likelihood of pervasive fraud and illegal voting. Meanwhile, ProPublica and other news organizations, alongside the Facebook Civil Rights Audit, have repeatedly drawn attention to the pervasiveness of misinformation about voting on Facebook. Despite a number of updated policies and efforts, including Facebook creating a dedicated site focused on reliable information about voting, we agreed the platforms can, and should, go further if they want to protect the 2020 US presidential election.
Our group came to agreement on four additional actions that Facebook, Twitter, and other platforms could take immediately to protect voting in the U.S. and around the world. These are necessarily limited by the timeframe of what can be accomplished between now and November. They are also complicated by the refusal of Facebook especially to release data necessary to analyze how users are engaging with content, disclaimers, and labels. And, these recommendations come in lieu of regulatory changes that have been continually deferred in the United States. However, taken together we believe that these changes would be important and beneficial modifications to how these platforms operate.
Misinformation about voting should be treated with the same urgency as misinformation about the U.S. census and COVID-19. The platforms have enforced their policies on the latter two against all users, including politicians. However, with the Facebook spokesman’s response “that the company won’t take action because of its long-standing policy of not fact-checking politicians,” the company keeps sending mixed messages about whether their electoral misinformation and voter suppression policies apply to everyone. There is a critical distinction between saying “we don’t fact check politicians” and “this post does not violate our policy.” The Facebook Civil Rights Audit highlighted this as one of its top concerns, writing “Facebook has made policy and enforcement choices that leave our election exposed to interference by the President and others who seek to use misinformation to sow confusion and suppress voting.” Facebook should clarify if it will, in fact, apply its voter suppression policies to all users, including politicians. Universally applied misinformation and voter suppression policies would also address the inconsistent ways in which politicians and other public figures are defined across platforms.
Facebook and Twitter should reconsider their labeling approaches. In the past year, Twitter, and more recently Facebook, began labeling false claims about voting by elected officials with generic disclaimers such as “Get the facts about mail-in ballots” (Twitter) and “Get Voting Information” (Facebook). While these are efforts at counter-speech — providing reliable information from institutionally-verified sources to counter misinformation — the current approach is unclear and likely ineffective. It may not be clear to all users, for instance, whether Facebook and Twitter are applying these labels to content they have determined is false, and if so, which assertions are false. The labels also lack any direct refutation of misinformation — both companies’ labels open onto generalized information about voting procedures, nothing that challenges the specific claims in these posts. Correcting false information at the point of contact (and not sending users elsewhere to figure it out) is especially important. This is one area where more available data from the platforms would help us analyze the efficacy of these policies. While we still would prefer the platforms to fact check all content related to voting and either remove or accurately label misinformation, if the platforms continue with these current policies, we strongly recommend accompanying them with the next two recommendations.
Facebook and Twitter should surface constructive counter-speech that elevates reliable information and place it directly throughout users’ feeds, not just when they click through on a label and not just at the top of a feed. In essence, they should create what we see as reliable ‘ambient election information.’ Most users scroll through social media, but we believe it to be highly unlikely that a significant percentage of users click through for more information when voting information is posted at the top of the feed or when a label is appended to a post. As such, reliable voting information will likely reach a negligible percentage of users compared to those confronted with misleading posts about voting. These companies should surface reliable electoral information regularly in the feeds of users alongside other content. This is about amplifying reliable information by pushing it out to voters, not requiring voters to seek it out for themselves.
To help amplify reliable information, Facebook and Twitter could provide resources (in the form of advice, training, content creation, design, ad credits, or free boosted content) to secretaries of state, boards of elections, local news stations, and other sources of trustable, reliable information, similar to what these companies routinely do for their commercial and political clients. These entities are bound by law to secure free and fair elections and administer voting. At the same time, they often have limited resources for communications and knowledge bases for engaging in communications and marketing on social media. As such, Facebook and Twitter can potentially play an important role in making their expertise more widely available to these state and local government entities to help them craft more effective content for voters. This might include helping them create better-performing organic content with details on voting registration processes or digital ads that refute election misinformation.
Moderating digital misinformation relating to electoral processes is deeply complicated but critical for the health of our democracy. While we remain concerned by Facebook’s inconsistent enforcement of its policies, we do believe that these strategies provide some options for these companies to push back against the manipulation of their platforms by those seeking to sow distrust in or disrupt the elections, while also avoiding charges of censorship. And this doesn’t end on November 3rd. Trustworthy information on the process of counting mail votes and certifying elections will be especially critical in the days, or even weeks, after November 3rd.
Stewart M. Coles, Illinois Distinguished Postdoctoral Research Associate, Department of Communication, University of Illinois at Urbana-Champaign
Yael Eisenstat, Visiting Fellow, Cornell Tech, Digital Life Initiative
Ellen P. Goodman, Professor, Rutgers Law School, Co-Director, Rutgers Institute for Information Policy & Law (RIIPL)
Young Mie Kim, Professor, School of Journalism and Mass Communication, Department of Political Science (Faculty Affiliate), University of Wisconsin-Madison
Karen Kornbluh Director, Digital Innovation and Democracy Initiative and Senior Fellow, German Marshall Fund of the U.S.
Daniel Kreiss, Principal Researcher, UNC Center for Information, Technology, and Public Life
Shannon C McGregor, Assistant Professor UNC Hussman School of Journalism and Media, Senior Researcher, UNC Center for Information, Technology, and Public Life
Eli Pariser, Co-Director, Civic Signals
Kathryn Peters, Executive Director, UNC Center for Information, Technology, and Public Life
Peter Romer-Friedman, Principal, Gupta Wessler PLLC
Francesca Tripodi, Assistant Professor UNC School of Information & Library Science, Senior Researcher, UNC Center for Information, Technology, and Public Life