Future Tense

This Is What’s Wrong With Twitter’s New “Manipulated Media” Label

Hands hold up a phone, with the thumb scrolling through Twitter.
Marten Bjork/Unsplash

On Saturday, Dan Scavino, the White House social media director, tweeted a 13-second clip of Joe Biden giving a speech in Kansas City, Missouri. In the video, Biden—“Sleepy Joe,” as the tweet calls him—stammers, then says: “Excuse me. We can only reelect Donald Trump.” The clip cuts Biden off midsentence, ending before he goes on to say, “if in fact we get engaged in this circular firing squad here.” The timing is perfectly calculated: Biden’s call for party unity looks, in the clip, like a confused endorsement of Trump.

Trump retweeted the video. Then, as the tweet was making the rounds Sunday, a baby blue warning appeared beneath it.

A screenshot of Scavino's tweet, with a still shot of Biden speaking, and the "Manipulated media" warning beneath.

This is the first time Twitter has used its “manipulated media” label, the Washington Post reported, since it was introduced last Thursday. The label is part of Twitter’s new synthetic and manipulated media policy, which the company announced on Feb. 4 amid growing concern over fake news and deepfakes, or A.I.-generated fake images and videos that look highly realistic, this election cycle. The policy states that Twitter may label tweets that are “significantly and deceptively altered or fabricated” or “shared in a deceptive manner,” and may remove tweets that are likely to “impact public safety or cause serious harm.” The policy also establishes criteria for determining a user’s intent to deceive.

After the policy was announced, critics expressed concern that it would set too high a bar for removing and flagging content. Fabricated media, they worried, would fall through the cracks.

At first glance, the application of the “manipulated media” label to the Biden video seems promising: It shows that Twitter is willing to clamp down on a more nuanced and mundane form of deception—in this case, lying by omission. (Scavino has since argued that it was “NOT manipulated” since he didn’t technically edit it, and his supporters are up in arms over Twitter supposedly setting a “dangerous precedent,” whereby all “clips” could be considered manipulated.) However, there are already serious concerns about the label’s function—and, more importantly, its design.

First, if a user searches for the tweet on Scavino’s profile or accesses it via a direct link, the label is missing. It’s currently only visible on individuals’ Twitter timelines and through the search bar. This is a technical issue that the company is working to resolve, a Twitter spokeswoman told the Washington Post on Monday.

More criticism has come on the design front. “Nice step but too subtle,” tweeted Dave Schlafman, a Netflix design director. “As a designer, @Twitter should make that label red so there’s no chance anyone could miss it.” He also pointed out that, in using its brand color, Twitter is making it easier for users’ eyes to move right past it. It’s also written in the same font (and font size) as the text of a tweet. Jesse Damiani, the deputy director of emerging technology at Southern New Hampshire University, agreed with Schlafman.  “It has to be a notable and undesirable badge,” he tweeted on Monday.

For now, that badge is a faint P.S. that almost blends in with hashtags and @s, both of which are also blue, when you scroll through your timeline. Twitter users have already called for the company to address this by using additional techniques, such as all-caps, better placement, and prominent watermarks.

Future Tense is a partnership of Slate, New America, and Arizona State University that examines emerging technologies, public policy, and society.