On Wednesday, the White House tweeted a link to a new survey asking participants if they felt they’d been discriminated against for their views online. “Social media platforms should advance freedom of speech. Yet too many Americans have seen their accounts suspended, banned, or fraudulently reported for unclear ‘violations’ or users policies,” the survey explains. “No matter your views, if you suspect political bias caused such an action to be taken against you, share your story with President Trump.” The Trump administration has made it clear it believes social platforms like Facebook and Twitter allow left-leaning users to post freely, while conservative thinkers’ posts are subject to unfair treatment not only from other users, but also from the networks themselves.
It’s no surprise that Trump would want to prove this allegedly biased treatment, though the survey’s vague questions and subjectivity might make it difficult to turn the responses into a quantifiable report. (There is also the fact that the private companies Trump has complained about have complete authority of who does and does not use their services; if Facebook wants to boot people off without cause, it can.) Regardless of its efficacy, the survey is yet another indication of the president’s complicated relationship with social media platforms.
What’s more eyebrow-raising than the survey itself is the user agreement participants are required to accept to have their answers taken into account. After submitting your name, ZIP code, phone number, and email address, hitting the “I Accept” button means survey-takers give the White House permission to reuse and edit their responses to any degree. “I think the fact that they want to edit it and they don’t want to ask your permission for the final form is kind of telling,” says Katharine Trendacosta, a policy analyst at the Electronic Frontier Foundation (EFF). The EFF actually created a similar survey tool that it has been using for years to collect and analyze stories of online censorship. Of course, its user agreement is on the same page and easily viewable, whereas the White House survey’s user agreement is available only by clicking outside of the survey page itself. The EFF’s survey also has something the White House’s does not: a mission statement. The organization uses the data provided to “shine a light on why companies make certain decisions about content, and how content takedowns are affecting communities of users around the world.” (Trendacosta points out that their survey also has a robust privacy policy, while the White House’s does not.)
While Facebook is partially the source of the White House’s agitation, the Trump administration would be wise to look at the social network’s own missteps with sweeping user agreements. In Facebook’s case, that strategy resulted in a lawsuit: Facebook was found to have falsely implied users “liked” things and showed them to other users in advertisements. The social network also agreed to a $20 million settlement following a class-action lawsuit over Sponsored Stories, a feature that allowed advertisers to turn “likes” into endorsements. As a result, Sponsored Stories were changed so that users had to opt in for their content to be used, and in 2014 they were killed altogether.
David Schultz, a political science and law professor at Hamline University in Minnesota, says the language of the White House survey user agreement is overly broad. “I understand the Freedom of Information laws and open government laws generally make any communication on government websites public information,” he says. “However there is a big difference between saying information you communicate versus saying you turn over all of your rights to let the government use your comments any way it pleases.” Schultz suggests that the way the user agreement reads right now, the White House can take responses and edit them to present positions that are opposite of what was originally intended. “So they could take this information of yours, they then twist it around, and they use it and basically say, ‘Well, so-and-so actually said glowing things about us.’” Schultz likens this to the classic review-blurb tactic: “Someone would review a movie and say, ‘It’s the greatest travesty of a movie of all time,’ and then, the movie advertiser would put the ellipsis out there, ‘The greatest … movie of all time!’” The potential for such tampering could make it difficult to trust the survey results at all.
The broadness of the user agreement raises a question: Why does the White House need such control over the presentation and possible reuse of the submitted answers? For private companies, it makes sense to employ user responses in advertisements or marketing campaigns, but what is Trump’s play here? The Constitution states that public resources can’t be used toward political campaigns, so if highly edited (or even unedited) responses from the survey end up in Trump’s 2020 campaign marketing materials, it would violate that law. Schultz also says that if a constituent’s response were used in some way they didn’t approve of, they could have grounds for taking it to court. “You could say, ‘Well, no, I never consented to that because you never told us what you were going to do with it,’ and you would probably prevail in court where a judge would strike it down and say they didn’t give you sufficient informed consent.”
The larger issue is that the survey will likely court biased responses. Surveys often do. Trendacosta says the EFF is aware of this issue in its own online censorship survey. “People who submit stories about their content being taken down are angry,” she says. She also points out that groups often marginalized and threatened by the government might be less likely to respond at all because they don’t feel safe offering their information. “So you end up with a very self-selecting group of people telling their stories, which doesn’t give an accurate or full picture of the problem of content moderation or online censorship or any of the sort of buzzwords around this topic,” she says.
When companies use this sort of language, it’s concerning, but not terribly unusual. In the White House’s case, it’s more alarming.The best course of action is to refrain from taking the survey altogether; self-censorship seems safer than agreeing to the White House’s conditions.
