Introduction
Unfiltered by the press, President Roosevelt spoke to Americans as if having an intimate conversation in their homes during his fireside chats.1 It was the first time a president used the new cutting-edge medium of radio in such a way, and it was effective, reaching millions of Americans at once.2 President Kennedy similarly pioneered the use of television to control his image and bypass conventional media, speaking practically face-to-face3 with tens of millions of his constituents.4 Understanding and embracing modern media technology is still a hallmark of presidencies.5 Although President Trump makes unprecedented use of Twitter,6 President Obama was the first president of the social media age,7 integrating digital technology into his administration8 to reach the 69% of Americans using those platforms.9
Hours into Obama’s presidency, he launched a website to circumvent established news media outlets and update the public about the administration directly from his team.10 By the end of his presidency, the President could be found on Facebook, Twitter, Instagram, YouTube, Snapchat, and Medium.11 His newly created Office of Digital Strategy managed these platforms with more staff than his Office of the Press Secretary.12 However, his goal to transform how the White House communicated with the public did not end with personal accounts.13 President Obama directed the entire executive branch to adopt social media practices with his Open Government Initiative.14 It aimed to improve transparency and public participation in rulemaking processes by providing more information to the public online.15 Today, every agency has at least tried using social media, but most are well-versed in navigating these platforms.16
While an increased use of social media has its advantages,17 this record push to bring agencies online also increased the risk of violating restrictions placed on agencies when communicating with the public.18 The Environmental Protection Agency (EPA) recently discovered how thin that line is. The agency is a pioneer of social media use,19 but stirred up controversy with its 2014 campaign to inform citizens of its proposed Clean Water Rule.20 The EPA used the more adventurous medium, Thunderclap, a platform that allows supporters of a message to broadcast that message at the same time, increasing awareness.21 The EPA shared the message “Clean water is important to me. I support EPA’s efforts to protect it for my health, my family, and my community,”22 which was posted on supporters’ feeds and thus spread to their followers.23 While this may not have turned the average user’s head,24 the Government Accountability Office (GAO)25 found that the EPA engaged in statutorily prohibited “propaganda” because nothing in the message itself indicated that the EPA wrote it.26 The GAO reads the statutory bar on propaganda to apply only to covert propaganda, meaning that the recipient of the message cannot identify the agency and its source.27 However, the EPA maintains that it did not violate any law because it posted the message on the agency’s account and took no steps to hide its identity.28
While agency use of social media is no longer innovative, questions remain about when those activities go too far. This Note proceeds in three parts. Part I observes statutory limitations on propaganda and the GAO’s interpretation of them. This Part explores why agencies use social media and analyzes past GAO opinions. Part II discusses the EPA’s Thunderclap campaign, evaluating both the GAO and the EPA’s arguments about whether the alleged EPA violation is warranted. This Part also notes how agencies have used social media after the GAO’s finding of the EPA’s violation.
Part III offers a judgment on which agency is correct, concluding that the EPA violated propaganda restrictions under current interpretations.29 However, the EPA’s contentions expose the inadequacy of the GAO’s approach to social media.30 What works with television or print media does not effectively balance the need to prevent propaganda activity with the valid interests of agencies to engage with the public through digital technologies.31 Lastly, this Part advises the GAO or Congress to define propaganda and proposes a definition. This Note argues that the GAO’s current practice is insufficient to ascertain whether a communication is propaganda because of its lack of attention to the actual content of messages. As is, harmless messages may be targeted while misinformation spreads so long as its authorship is clear.
I. Background
A. Congressional Limitations on Agency Outreach
Congress has long been concerned with agency communication with the public.32 Doing so allows control over agency conduct, priorities, and messages ensuring consistency with Congress’s goals.33 For example, the Federal Anti-Lobbying Act prevents grassroots lobbying by a federal agency by prohibiting payment to influence any government official to vote a certain way on any law or policy.34 Similarly, through a joint regulation implementing a congressional bar, multiple federal agencies prohibit themselves from paying a person to influence a member or employee of Congress.35 Read broadly, this prohibition covers any action taken by agency employees directed towards Congress, as those employees are paid by an agency.36 The Federal Antideficiency Act forbids any government officer or employee from permitting an expenditure not authorized by Congress or in excess of the amount appropriated by Congress for such spending.37
Each year, Congress passes an appropriations statute and, in almost every year since 1951,38 forbids any of the appropriated funds to be used for “publicity or propaganda.”39 Notably, this provision does not define “publicity or propaganda.”40 Indeed, Congress has not provided a definition since the phrase first appeared, when Congress voiced its unease about a campaign to promote a national healthcare plan.41 During the mid–to–late–1940s, politicians feared that the healthcare plan was socialist and that promotion of it constituted impermissible propaganda.42 They used the appropriations law language, which prohibited using funds for propaganda, to defeat the plan.43 When questioned about a definition of propaganda, the Congressman who introduced the language simply responded: “We can well distinguish between what is propaganda and what is educational matter.”44
Enforcement of this limitation falls to the GAO.45 The GAO is an independent agency.46 At Congress’s request, it investigates how agencies used their budgets for a specific activity.47 The GAO must balance the need to carry out Congress’s will with an agency’s right to inform the public of its activities, which includes justifications and defenses of those activities.48 The GAO interprets the term propaganda, and thus the scope of impermissible propaganda, as covert propaganda only.49 Although the GAO has not provided a definition of “propaganda,” it does describe what it considers to be covert propaganda: messages whose authorship is unclear such that the agency’s role in broadcasting the message is not disclosed to the public.50 The GAO’s determination of whether an agency material is propaganda does not focus on the content of the message, but on the clarity of the agency’s role in disseminating such material.51
B. Agency Use of Social Media
Since the White House live-streamed its annual Easter Egg Roll in 2002, government agencies have increasingly used social media to the point where it is commonplace today.52 Agencies are found across the social media spectrum, using any platform available to engage with the public.53 The EPA alone runs a website, thirty-seven Twitter accounts, nine blogs, a YouTube channel, Pinterest, Instagram, Flickr accounts,54 and more.55 Similarly, the Department of Agriculture (USDA) has twenty-nine Twitter accounts,56 and the Department of Education (DOE) has over fifteen Twitter accounts, ten Facebook pages, five YouTube channels, and multiple blogs and Storify accounts.57 Thanks to the widespread belief that social media is the new public square, allowing for greater engagement with citizens, government agencies recognize that social media platforms offer a wide reach to new audiences and, whether successful or not,58 dedicate a lot of time and money to cultivate a real presence online.59
Agencies conduct social media campaigns for different reasons, from raising awareness of an issue to requesting feedback.60 For example, in 2016, the Department of Labor began a campaign to raise the minimum wage.61 The campaign included a web page dedicated to the issue, videos on YouTube, and a Twitter hashtag “#RaiseTheWage.”62 Similarly, the USDA’s Forest Service recently revived the World War II-era mascot “Smokey Bear” to inform the public about wildfires.63 Using Facebook, Instagram, and Twitter accounts, the USDA created a “softer image” of Smokey Bear for the modern audience.64 On these outlets, the USDA featured Smokey Bear on artwork with phrases such as, “Make it Your Goal to Extinguish Hot Coals.”65 Although a lighthearted campaign, an estimated 6.7 million acres of forests are destroyed by fire yearly,66 with nine out of ten wildfires started by people.67 Using social media, the USDA is able to inform a modern audience of serious issues.68
With agencies increasingly using social media to communicate, there is much to monitor.69 As this Note discusses in Part III, treating social media messages in the same manner as those issued through traditional media can negatively restrict agencies and prevent any valuable outcomes of using social media from ever coming to fruition.70 The GAO is consequently tasked with navigating a complicated but important social media landscape, and has responded by making reports on the challenges of using such platforms.71
C. GAO Rulings on Covert Propaganda
Congress has prohibited certain agency communications through its appropriations laws, but it has not set up a robust scheme of enforcement.72 Rather, because the anti-propaganda provision is an appropriations measure, enforcement lies entirely with the GAO.73 Individual members of Congress ask the GAO to conduct investigations into specific agency actions.74 The purpose of the investigation is to determine whether or not the agency used appropriated funds for prohibited purposes.75 Thus, a finding of propaganda activity is a conclusion that the agency spent taxpayer dollars unlawfully.76 However, the significance of a GAO finding is not that legal conclusion, but the political consequences. If the GAO labels an agency campaign as propaganda, members of Congress—particularly those who are not from the president’s party—use that conclusion to promote opposition towards that campaign.77 This Section analyzes some of the GAO’s reports on agency activity to ascertain what covert propaganda means in practice and what the GAO considers when making a judgment.
1. Finding No Violation
The GAO typically defers to agency justifications for its spending.78 For instance, early in the Obama administration, during an intense public debate over what ultimately became the Affordable Care Act, the Department of Health and Human Services (HHS) posted electronic form letters on its website for the public to sign, which contained declarations such as, “[w]e strongly support your commitment to comprehensive health reform.”79 The HHS maintained that its only purpose was to poll public opinion.80 Because the letters were posted directly on the agency website, the GAO found that the HHS clearly identified itself, and therefore the letters were not covert propaganda.81 The GAO stressed that it generally defers to agency discretion on how they inform the public.82 Unless the justifications for such activities are obviously flawed and unreasonable, the GAO will not intervene.83
In 2009, the Department of Defense (DOD) established a retired military officers outreach program, consisting of regular meetings with retired officers about the war on terrorism and paid travel to various military locations.84 The DOD created the program because it knew those officers were involved in public outreach and would spread a DOD-positive message.85 The GAO admitted that the DOD was clearly trying to influence public opinion of its war policies through the officers, but determined that this type of subject matter did not violate propaganda prohibitions.86 Regardless of the goal, because the DOD did not hide its identity when providing the officers with information or travel, its actions did not amount to covert propaganda.87 Any promotions made by the officers after those meetings were not paid for by the DOD, and therefore would not fall under any appropriations statute.88
The most recent GAO finding that agency communication was lawful involved Twitter.89 Although a Twitter hashtag campaign may seem susceptible to the spread of false information because any user can use an agency’s hashtag in their own message,90 the GAO held that the EPA’s hashtag campaign did not amount to covert propaganda.91 Along with ruling on the EPA’s Thunderclap campaign in 2015,92 the GAO considered whether the agency’s #DitchtheMyth campaign, designed to combat against a #DitchtheRule campaign expressing opposition to the EPA’s proposed rulemaking change,93 violated the covert propaganda prohibition.94 On its website, the EPA posted prewritten tweets for users to share on their own accounts, containing the hashtag #DitchtheMyth.95 The GAO ultimately concluded that the EPA activity was permissible because, while the campaign allowed users to post messages written by the EPA from their accounts, all graphics the EPA posted had the agency logo and prewritten tweets included the EPA’s Twitter handle at the end.96 These inclusions indicated to the EPA’s intended audience that the EPA was the source of any information publicized.97 As always, the GAO was solely concerned with whether the message was covert, not with whether its content constituted propaganda. The GAO found that so long as any tweet written by an agency can be identified as such—by being posted from an official account or including the agency’s Twitter username in the message—starting a hashtag campaign is permissible.98
2. Finding Violations
The GAO’s decision on the EPA’s #DitchtheMyth and Thunderclap campaigns was the first time it considered whether social media activity violated appropriations law.99 It is helpful to begin with GAO findings of a violation on older media, because the GAO’s considerations remain the same for new platforms.100 In 1986, the GAO investigated the Small Business Administration’s (SBA) preparation of suggested editorials for distribution in newspapers.101 The editorials were letters to government officials offering support from a third-person perspective for the Reagan administration’s effort to move the SBA to the Commerce Department.102 The GAO found that the agency did not simply disseminate information, which is a legitimate activity.103 When published, the editorials appeared as if the views it promoted were those of the newspaper, not of the agency itself.104 The agency’s authorship was concealed so the public would not know that the SBA used its money to push that support.105 The GAO found the editorials to be propaganda because they were deceptive as to their authorship.106 The GAO interpreted Congress’s prohibition on propaganda to be only a prohibition on misleading the role of the agency in its messages.107 An agency may still promote its ideas, garner support, and advocate a perspective without violating propaganda restrictions, so long as its role in dissemination is unambiguous.108
In 2004, the HHS hired a public relations firm to release video news releases (VNRs), or prepackaged news reports,109 on television about new laws designed to improve Medicare coverage.110 They included a script for a broadcaster’s news anchor to read when announcing the VNR that read as if the anchor were introducing a real news segment and did not identify that it was pre-made by the HHS.111 With a fake news anchor reporting, the VNRs contained only facts about the new law but did not indicate that the videos were produced by the government.112
While the HHS maintained, and the GAO recognized, that using VNRs is standard practice for government entities and the news sector, the GAO took issue because the videos were misleading as to their source, which meant they constituted covert propaganda.113 While the news station itself may have known the agency paid to create the VNRs, the HHS’s target audience—the public—would believe the information came from a neutral source and not the government.114 The HHS’s overall mission and message may have been innocent, but concealing itself as the source of the videos pushed the message into unauthorized territory.115 Conversely, the Department of Justice’s Office of Legal Counsel disagreed with the GAO in this case, finding that, because the VNRs did not advocate for a particular viewpoint or opinion and instead were purely informational, they did not constitute covert propaganda simply because the HHS’s role in their production was not revealed.116
The GAO continues to focus on clarity of source rather than on content of messages, as evinced by its 2005 investigation of the DOE’s No Child Left Behind Act advertisements.117 The DOE retained Armstrong Williams, founder of a public relations firm, to regularly comment on the Act during television and radio broadcasts in order to educate minorities about the new law.118 Williams commented on the Act over one hundred times without disclosing his relationship with the DOE,119 and his contract did not require that he disclose to any audience that he was being paid for his statements.120 Therefore, the DOE concealed its role, making it appear as if an independent source advocated for the agency’s position, when all the while such sponsorship was in exchange for public funds.121 The GAO found that the DOE’s conduct constituted covert propaganda.122 Notably, the GAO asked the DOE for a transcript of Williams’ activities not to analyze the content of the messages, but to see whether or not he made relevant disclosures.123
Just four years later, the HHS faced a situation almost identical to the DOE’s No Child Left Behind Act incident when it contracted with an economist to promote a health care plan.124 The economist did not disclose his financial relationship with the HHS in his editorials.125 The GAO did not rule on this, but would likely have been as concerned, as it was with the DOE, that the HHS did not ensure that this paid representative reveal to the public his connection to the agency.126 The GAO would probably consider the economist’s communications to be covert propaganda.127
These cases show that, in finding an agency action constitutes covert propaganda, the GAO strictly considers whether audiences would realize a message was broadcasted by an agency. While the GAO defers to agency judgment, it applies a test to determine whether a communication was covert—whether it is clear an agency was the author of its message—and does not look to what was in the communication itself. In the 2009 DOD case, the GAO found that the subject of the agency’s messages to retired military officers was not a concern for a finding of propaganda, despite the agency’s goal of using those officers to garner support for itself.128 Conversely, the GAO found that the HHS’s VNRs constituted covert propaganda even though the HHS was in an analogous position to the news station as the DOD was to the retired officers.129 The two are distinguishable in that the HHS’s intended audience was the viewers watching the news.130 The DOD shared messages about itself with no guarantee that the officers would spread those messages.131 Like the DOD case however, the GAO did not look at the content of the HHS message to establish whether it was propaganda.132 There were no inquiries into whether the communications were advocacy, or whether any facts asserted were accurate.133 Similarly, the GAO did not analyze what the EPA’s #DitchtheMyth tweets said, asking only whether they were properly credited to the agency.134 A covert propaganda investigation determines whether an agency disclosed itself as the source of a communication to its intended audience.135
Although not investigated by the GAO, Congress’s Committee on Oversight and Government Reform reported on the Obama administration’s first year in office, detailing instances that likely would have violated the GAO’s covert propaganda restriction.136 One in particular dealt with blogging websites.137 The Office of Public Affairs138 hired Tracy Russo to serve as its New Media Specialist.139 Congress’s report claims140 that Russo deliberately searched for online editorials criticizing the administration and anonymously posted comments countering any negative arguments.141 This type of persuasion of the public would not raise covert propaganda concerns had Russo’s employee relationship with the Office of Public Affairs been disclosed in her comments.142 But because the posts were anonymous, the GAO would find that the Office of Public Affairs violated the propaganda prohibition.143 The GAO’s covert propaganda test remained the same since entering the social media realm, as exemplified by its ruling on #DitchtheMyth,144 but that test may have unintended effects as agencies struggle to apply old rules to a new medium.145
II. Analysis
A. Events Leading Up to the EPA’s Thunderclap Campaign
The Clean Water Act (CWA) aims to protect water quality and reduce pollution.146 It does so primarily by limiting discharges of pollutants into “navigable waters,” which are defined broadly as “waters of the United States” (WOTUS).147 But the CWA does not define that phrase, and the protections and impact of the Act depend on what is considered WOTUS.148 Pollution of a water body that is not a WOTUS is not a violation of the CWA, and a narrower meaning limits EPA jurisdiction while a broader one expands it.149 The Supreme Court sought to clarify the term in 2006 in Rapanos v. United States,150 focusing on whether wetlands were covered.151 However, the case resulted in a 4-1-4 split decision, leading to more debate.152 Justice Scalia, writing for the plurality, advocated for a restrictive definition of WOTUS, holding that the phrase includes only waters adjacent to a navigable waterway.153 Justice Kennedy’s concurring opinion was broader, and mushier, defining WOTUS as any body of water that has a “significant nexus” to a traditional navigable waterway.154 The remaining four Justices argued that the EPA can regulate much more: any body of water, including non-navigable ones, that is hydrologically connected to navigable waters.155
Rapanos created a confusing situation, and circuit splits across the nation, and the EPA was left to choose which of the three Supreme Court Justices’ opinions to follow.156 Under the Obama administration, the EPA undertook this task and made some headway in 2011 with proposed guidance using Justice Kennedy’s significant nexus test.157 The EPA, with the Army Corps of Engineers, formally proposed its Clean Water Rule in 2014.158 By the time the comment period ended in November of 2014, the agency had received over one million comments, 87% of which were in support.159 In 2015, the EPA published a final rule, essentially codifying Justice Kennedy’s significant nexus test.160 Before finalizing the rule, however, it undertook a social media campaign to promote and inform citizens of the Clean Water Rule, as well as to counter online opposition.161 While the EPA was still accepting comments to the Rule, the American Farm Bureau started a Twitter hashtag162 encouraging the public to #DitchtheRule because it believed the EPA was expanding its jurisdiction beyond its constitutional reach.163 Over 230 members of Congress supported the Farm Bureau’s efforts.164 The EPA sought to respond to the Bureau’s “facts”165 and advocate its belief that the Clean Water Rule would prevent water pollution while bypassing traditional media.166 In addition to its #DitchtheMyth Twitter campaign, the EPA took to Facebook, YouTube, and the new (even for social media) platform Thunderclap.167
B. The EPA’s “I Choose Clean Water” Message
Thunderclap considers itself a social media “flash mob” or, more practically, a massive scheduled social post.168 A user can start a campaign by creating a Thunderclap message and asking others to “support” it or share it, which schedules the message to automatically post on the supporters’ other social media accounts at one time.169 At that scheduled time, the message is displayed on every supporter’s social media account, spreading the campaign across multiple websites and reaching a larger audience than one would by posting the message on, for example, Twitter alone.170 The EPA acknowledges that a primary reason for using Thunderclap is reaching that secondary audience: the supporters’ social media followers.171
Titled “I Choose Clean Water,” the EPA’s Thunderclap campaign message to be posted on supporters’ social media feeds read: “Clean water is important to me. I support EPA’s efforts to protect it for my health, my family, and my community.”172 It ended with a URL, the EPA’s “message link” that Thunderclap requires users include with a message, to direct supporters to an outside site where they can learn more about the campaign.173 The link led to the EPA’s official web page on the Clean Water Rule.174 The message had 982 supporters and reached an audience of over 1.8 million people.175
1. The GAO’s Response
Republican Senator Jim Inhofe, later aided by two other Republican senators, asked the GAO to investigate the EPA’s Thunderclap campaign.176 Concerned with how the secondary, anticipated audience would know of the agency’s authorship, the GAO found that the campaign constituted covert propaganda.177 The GAO noted that the campaign’s page clearly indicated that the EPA was its creator,178 but found that the message itself did not indicate that the EPA wrote it.179 This is a problem because once the message was posted on the supporters’ accounts, the secondary audience viewing the message from those accounts would not be able to tell that the EPA originally wrote the message.180 Comparing this situation with the HHS’s VNRs, the GAO emphasized that it is not enough for an agency to divulge its authorship to the conduit of messages if it is concealed from the target audience.181 The HHS’s notification to news stations was not enough because its intended audience, the public, would not know of the HHS’s role.182 Similarly, the EPA’s disclosure of its identity to the supporters of its campaign was not enough because its intended audience, the secondary users, would not know of the EPA’s role.183
The GAO also found that the first-person point of view of the message furthered concealment of the EPA’s authorship, stating that the agency purposefully made it appear as if it were not the writer.184 It would appear as if a supporter wrote the message herself once it was shared on her account, without any ascription to counter that assumption.185 The GAO found the message retained no identifying information once shared,186 violating the statutory restrictions on using funds for propaganda purposes.187 The GAO asked the EPA to determine the exact amount it cost the agency to conduct the Thunderclap campaign, and to repay that amount to the Treasury.188 White House officials were silent as to whether they agreed that the EPA conducted propaganda activity.189
2. The EPA’s Defense
The EPA disagrees with the GAO’s outcome.190 The agency reaffirmed that in this age, social media is one of the most effective tools to share information and be transparent with the public.191 The campaign aimed to get citizens engaged with the rulemaking process, which is a normal use of the platform.192 The EPA claims that it made no effort to hide its ownership of the Thunderclap because the agency published the message on its own, clearly labeled account and linked to its official website.193 The EPA thinks it important that users chose to support the message, but focusing solely on the Thunderclap’s supporters avoids a major issue the GAO had with the campaign.194 For instance, the EPA makes no mention of the message’s point of view, nor of the fact that the EPA was actually targeting the secondary audience.195 The GAO believes that the message’s supporters were merely a conduit to the agency’s intended audience, making the EPA’s defense of how it appeared to the initial supporters (rather than the defense of how it appeared to the secondary audience) largely irrelevant.196
The EPA later rejected the claim that the campaign supporters were only conduits of the Thunderclap message, emphasizing again that the supporters had the freedom to choose whether to put the message on their accounts, and even had the opportunity to edit the message.197 The EPA likens this situation to the GAO’s ruling on the DOD’s retired military officer outreach, where the DOD did not pay the officers but knew they would spread a pro-DOD policy message to the public.198 The EPA argues that, like the DOD, it had no control over the message that the Thunderclap supporters would actually post.199
Regardless of whether the EPA engaged in illegal activity, the GAO’s penalty is not the most effective deterrent to stop agencies from undergoing propagandist campaigns in the future. The GAO’s authority to penalize the agency is limited to requiring that the EPA determine how much money was spent on the Thunderclap campaign.200 However, while many find the GAO to be largely ineffective in its solutions,201 the political effects of the GAO’s finding are a greater consequence for the EPA.202 The GAO’s ruling played into Republican and conservatives’ opposition to the rule, and to the EPA itself, providing challengers with the ammunition they sought to discredit the agency and its plan.203
C. Recent Agency Social Media Use
After the GAO’s 2015 report on the EPA, the House passed the Regulatory Integrity Act, which requires the reporting of every agency communication about a rule and prohibits agencies from promoting a rule on social media.204 Agencies would not be allowed to advocate for a rule; they could only use social media to inform citizens of the content or status of a rule.205 This Act would prohibit more than covert propaganda as defined by GAO; for example, it would seem to prohibit even the EPA’s #DitchtheMyth campaign.206 Members of Congress are concerned that agencies use social media to sway public opinion, rather than sincerely invite criticism and feedback about rules.207 While some find the Act to be overly burdensome,208 the possibility of its passage, along with the GAO’s finding of propaganda, may be contributing to a chilling effect on agencies, which are not using social media as casually as they once were to inform the public about rulemakings.209 For example, in 2016, the EPA only tweeted once about a public comment period, while the Department of the Interior’s and Bureau of Land Management’s Twitters featured zero posts about public comment periods.210
In a role reversal of the Thunderclap incident, Democratic members of Congress asked the GAO to review whether the EPA under the Trump administration211 violated the propaganda prohibition.212 EPA Administrator Scott Pruitt appeared in a promotional video produced by the National Cattlemen’s Beef Association to oppose the 2015 Clean Water Rule.213 He spoke about why he is against the Rule, and urged farmers to provide comments on the proposed repeal of it.214 The video directs viewers to BeefUSA.org to file comments, which has the title “Take Action Now—Tell EPA to Kill WOTUS.”215 In the video, Pruitt also said that the Clean Water Rule defined WOTUS as including puddles, dry creek beds, and ephemeral drainage ditches, but the Rule specifically excludes puddles, dry creek beds without beds, and ephemeral ditches that only flow after precipitation.216 The GAO will investigate the situation,217 but will likely not look at whether those assertions were misleading or false.218
III. Proposal
A. The EPA Violated the Law
Under the GAO’s interpretation of annual appropriation bills, the EPA’s Thunderclap campaign was propaganda.219 Using the GAO’s test of considering attribution of a message only, the GAO is correct that the campaign’s first-person point of view makes it covert propaganda.220 Unlike providing plain facts, the EPA encouraged users to share a subjective message: “I support EPA’s efforts.”221 Because of its perspective, the EPA knew that once the message appeared on users’ profiles, it would appear as if the user wrote it themselves.222 The EPA is removed from the post, regardless of the fact that the agency posted it on its account.223 Although the EPA insists that its target audience was only the supporters, the EPA used Thunderclap specifically to target the supporters’ followers and reach a larger audience; the agency’s list of social media platforms that it utilizes explains that the EPA uses Thunderclap to have supporters promote agency messages to their social media followers.224 The perspective of the message thus makes the campaign amount to what the GAO considers covert propaganda.
What further distinguishes the Thunderclap campaign from other social media efforts is that it lacked identifying information within the content of the message.225 Thunderclap is a unique platform because, unlike a retweet226 on Twitter or a share227 on Facebook, the EPA’s authorship would not be reposted on supporters’ social media feeds when the message posted on their accounts.228 The only feature in the Thunderclap message that could be an ascription is the link concluding it, directing the public to the EPA’s official website.229 Neither the GAO nor the EPA focused on this part of the message.230 For the EPA’s #DitchtheMyth campaign, the agency added its campaign hashtag and Twitter username at the end of pre-written tweets to indicate that it was the source of the messages.231 In contrast, a link to the EPA’s official site does not as effectively disclose the message’s origin. A follow-up link to explain the EPA’s Clean Water Rule is not a clear ascription because people typically add links to their messages for their audiences to get further information about a topic.232 Intended innocently or not, the EPA’s Thunderclap hid the agency’s identity to its intended audience, making the message covert and violating the GAO’s interpretation of propaganda.233 However, actions that would satisfy the GAO may not be feasible.234 If part of the GAO’s goal is to protect an agency’s right to inform and engage the public, it must reconsider what covert propaganda means on the Internet.235
B. The GAO Needs to Consider the Nature of Social Media
The message that the GAO sent agencies is clear: an agency openly posting on its own account is not always enough; it must also include in the text or image of messages themselves a credit to the agency.236 The GAO compared its social media investigations to its decisions on agency activity through older media, trying to mold these violations into traditional ones and ignoring the characteristics and exceptional reach specific to new platforms.237 But what may have constituted covertness over television does not necessarily do the same on social media. Covertness may not be an effective standard to hold messages to in order to prevent propaganda.
For example, Thunderclap does not allow Twitter usernames in messages and encourages messages in the first person.238 In many ways, the EPA’s use of Thunderclap was a typical and natural application of the tool.239 In fact, DigitalGov240 directs agencies to use Thunderclap for issue or topic support,241 and the White House itself followed these norms when it made a Thunderclap campaign in 2013 with a message in the first-person point of view.242 Further, Thunderclap has a 140-character limit on messages;243 similarly stringent length limits are typical of social media.244 Shorter, clean messages are valued on social media, and an ascription or link after a message can take up significant space.245 On social media, agencies try to publicize a complex message in little space and cannot always fit an attribution that undoubtedly credits a message to them.246 Some believe the GAO went too far in finding that the EPA’s Thunderclap message was a violation, arguing that there was enough accreditation in the agency’s post and that the GAO’s concerns are trivial compared to other violations of the propaganda prohibition.247
Normally, an agency can avoid violating the GAO’s disclosure requirement because messages are usually posted from the agency’s official account.248 If users share an agency’s message, the medium will show that the agency is the author.249 On television, it may be easy for a paid government employee to appear on screen and hide their relationship to an agency, acting as if they are a nonpartisan citizen.250 Online, sites try to build such verification into their design, so an employee posting from an agency’s official account can have her message shared without obscuring her identity.251 The trouble occurs in situations where agencies create prewritten messages or use platforms such as Thunderclap—where the audience is the first to share the agency’s message.252 However, as discussed, the traditional idea of covertness may not mean as much online as it does in older media. Therefore, the GAO should consider online norms when determining whether activity is propaganda. Because covertness alone is problematic in determining whether a communication is propaganda, the “congressional watchdog”253 must return to the statutory text and focus on what propaganda actually is.
C. The GAO or Congress Needs to Define Propaganda
When Congress first enacted prohibitions on agency “propaganda,” it was worried about the indoctrination of socialism.254 Some members of Congress successfully prevented the implementation of a national healthcare plan, calling the campaign to promote it propaganda and adding the term to its annual appropriations bill.255 While knowing which messages were paid for by the government would be an important interest to identifying propaganda, the healthcare plan in this instance was clearly proposed by the government.256 It was fear of socialists within the government and their manipulation of the public that spurred the addition to the appropriations bill.257 Not all propaganda is covert.258
An important question then remains: even if the EPA’s Thunderclap message was covert, was it propaganda? Neither the GAO nor Congress defines propaganda.259 Under the GAO’s longstanding approach, whether an agency is engaged in covert propaganda turns solely on whether it is apparent to the public that the agency is the source of the message; the GAO ignores the content of the message.260 While the transparency of whether the government paid for a message is vital for accountability purposes, the content should be as much of a concern.261
Americans increasingly rely on social media to get their news,262 and studies show that getting information online fosters political engagement.263 Digital technology may allow the manipulation of voters more easily than other media,264 and a message delivered by the government affects how the public perceives it, and thus affects its persuasiveness.265 In addition, Internet media amplifies messages faster and to a larger degree than other traditional media, making for very persuasive platforms.266 These are not inherently dangerous realities, but can make the public more susceptible to government propaganda.267 However, propaganda is not always associated with a negative connotation,268 and any definition of it for the GAO’s purposes must navigate around this idea. Merriam-Webster’s definition of propaganda is “the spreading of ideas, information, or rumor for the purpose of helping or injuring an institution, a cause, or a person.”269 That definition covers a broad range of activity, even the legal dissemination of messages today.270
Although the GAO maintains that it does not know of a satisfactory definition of propaganda that excludes permissible agency communication due to the lack of direction by Congress,271 either it or Congress needs to attempt to define it. Government propaganda does not indoctrinate Americans, especially on the Internet, where citizens can voice criticism and speak back to the government.272 But it can still be dangerous. Propaganda deceives people by appealing to emotions, relying on stereotypes, making false claims, contorting facts, and more.273 It is particularly operational in a polarized environment, which increasingly describes our nation, and can threaten democracy by distorting reality.274 This vulnerability is why the GAO should consider the content of an agency’s message.275 For example, while the GAO found that the EPA’s #DitchtheMyth campaign was not propaganda, the public remains unaware of whether the agency’s purported facts are actually true.276 Verifying information publicized by an agency, especially when that information is used to counter opposition and raise support for a pending rule, would satisfy a concern that is currently lacking in the GAO’s reports.
This is not to say that the GAO should abandon its current test. The GAO should continue asking whether an agency’s message is covert, but should also take into account the type of message being conveyed. The GAO believes that establishing a definition of propaganda is too problematic because of the impossibility of reconciling impermissible activity with an agency’s right to inform the public and defend its policies.277 But the GAO can create a definition by supplementing its current test with other factors. In addition to covertness, such other factors should include: (1) deceptiveness, which focuses on whether the content of a message is false or unusually exaggerated or misleading; and (2) emotivity, which focuses on whether an agency deliberately and egregiously plays on the public’s emotions. Therefore, propaganda would be any agency communication that is covert, deceptive, and/or egregiously emotive. If the GAO treats these three as distinct, the agency will find propagandistic messages in every investigation. For example, while most agencies present certain views in favorable or unfavorable lights,278 the GAO does not have to label that language as unlawful, misleading propaganda. Instead, the GAO should balance all three factors in relation to each other, treating them as prongs to the greater propaganda phrase. It should consider the context and medium of a message to determine whether it implicates the propaganda prohibition.
If the EPA’s Thunderclap campaign were analyzed under this test, the first-person perspective of the message would not weigh as heavily because of how Thunderclap is used. That the agency’s intended audience would not know the EPA wrote the message is a concern,279 but would be placed in a larger context. The EPA used Thunderclap the way it was directed to, the way the website suggests, and the way most use the platform.280 The GAO would then need to determine whether encouraging citizens to tell their social media followers that they support the EPA’s proposed rule is propaganda. There are no facts in the message, but it does appeal to the public’s emotions with simplistic ideas. Appealing to emotions alone cannot constitute propaganda because much of what the government does is sensitive and concerns health, family, and community. However, it is noteworthy that this message was posted in the midst of a proposed rulemaking, but, deceptively, does not indicate in the message itself that there is a rule at all.281 The message contains no indication for readers to know that the EPA’s efforts are for anything other than general goals of encouraging a safe environment.282 It is true that the EPA validly promoted its goal and mission for supporters to share, while providing a link for users to get more information on the EPA’s concrete plans.283 Still, because the EPA tried to sway the secondary audience’s opinion with such an usually misleading message, the Thunderclap campaign would still count as propaganda under the approach I propose.284
That the GAO’s report had the same outcome does not mean its current approach is effective for social media. There are instances in which the GAO analyzing a message’s covertness alone would not correspond with one where it considered a broader view of propaganda. If the EPA was more specific in its Thunderclap about what its purpose was,285 the message would be propaganda under the GAO’s current approach, but would not under a more flexible, multi-factored one. The latter approach understands that Thunderclap works differently than older media.286 This is crucial if the GAO does not want to discourage agency use of new platforms, which can reach modern audiences more effectively than ever before, by finding propaganda where most of the public would not.287 Additionally, if the EPA’s #DitchtheMyth purported facts were untrue, the GAO’s current approach would still not find propaganda, while the proposed definition would. This is because the definition takes into account how propaganda works beyond a message’s unclear origin, to better protect the public against all types of propaganda.
D. Possible Objections
Because federal agencies must engage in some degree of covertness on social media, it may be submitted that the GAO should discourage agency use of those platforms entirely. However, sharing messages is an essential feature to communicating and spreading information online.288 Simply because covertness may be inherent to some social media does not mean that agencies should stop using social media altogether. The advantages of social media are significant, allowing the government to communicate easily, directly, and at little cost to a large number of Americans at any time in ways that are not possible using older media.289 The GAO should therefore permit normal uses of social media to promote agency activity and encourage public participation in politics and the rulemaking process.290
The GAO’s primary concern in adopting a more comprehensive propaganda definition is that broadening it may make otherwise permissible government communications unlawful.291 Yet the difficulty in drawing that line should not dissuade the GAO from doing so. Propaganda may not brainwash Americans,292 but it does make citizens susceptible to manipulation.293 The GAO’s instinct to adopt a narrow interpretation of propaganda is reasonable, because making a broad definition may render most agency communication unlawful.294 However, this Note’s proposed definition requires the agency to balance covertness with deceptiveness and emotivity, finding violations when an agency lies to the public or effectively does so with egregious manipulation. Therefore, an agency would not violate the law when persuading campers to avoid accidentally causing forest fires,295 but may when purporting to tell the public essentially false facts about a rule.296 The current GAO approach is both underinclusive and overinclusive—while many types of suspect agency propaganda are ignored, arguably harmless communications are classified as covert propaganda.297 The proposed definition expands the GAO’s interpretation, but is still limited, and even reduces some classifications of covert propaganda.298 This proposed analysis may be more subjective than the GAO’s approach, but it does not give the GAO so much discretion so as to preclude a consideration of special factors.
Broadening the definition of propaganda may also raise the concern that it unnecessarily expands the amount of communication open to GAO investigation, burdening the agency.299 However, because the GAO does not affirmatively look at agency messages, it would only increase the number of requests it receives from Congress, which should not result in encumbering the GAO with investigations.300 As to the idea that having the GAO verify the truthfulness of every agency statement would be too burdensome, the GAO would likely, at the very least, only require an agency’s justification and sources for its statements; this is a practice the GAO currently employs to determine whether an agency violated propaganda prohibitions.301 Doing this would still encourage agencies to ensure that they can credibly support their messages, which would in turn better inform the public and help combat against common propagandistic practices.302 Relatedly, it may be suggested that expanding the definition will have a chilling effect that this proposal seeks to prevent: agencies refraining from using social media for fear that any message will be labeled propaganda due to its misleading or emotive nature.303 However, the proposed definition would still allow misleading or emotive messages. Additionally, agencies will likely not stop communicating online,304 but may focus more on the truthfulness of their messages—which some believe is a good idea.305
This Note’s definition of propaganda adds a focus on content only. Agencies will be more concerned with what they are posting, on any platform, rather than with where they are posting, unlike the effect the GAO’s approach has.306 Agencies should feel free to use all social media platforms, but should be more diligent in the type of message they broadcast. This closes the large gap that the GAO’s approach leaves, allowing other propaganda activity to remain unaccounted for.
Conclusion
With modern technology advancing at such a fast pace, social media techniques should not be evaluated the same way that the GAO looks at those on older media forms.307 Covertness in the GAO’s long-established sense of the word may be unavoidable at times, but the answer is not to effectively restrict agency use of social media altogether.308 Rather, the GAO needs to have a more adaptable approach that considers the type of medium a message is on and its norms.309 Further, either the GAO or Congress needs to define propaganda to allow an interpretation that covers the content of a message. This includes a focus on whether a message is deceptive or egregiously emotive.310 What may not be covert propaganda may be substantively propaganda, and vice versa.311 To better protect our democracy, the GAO must revisit its interpretation to ensure that agencies are held accountable for such unlawful communication with the public.312