Twitter on Sunday applied its “manipulative media” label to a deceptively edited video showing presidential hopeful Joe Biden saying, “re-elect Donald Trump.”
It was the first time the service enforced rules adopted last month to control synthetic and manipulated media.
The flagged video was posted to Twitter by Dan Scavino, White House director of social media, and retweeted by the president.
By the time Twitter labeled it, the video had been viewed 5 million times and retweeted 20,000 times, according to The New York Times.
Twitter’s policy states it will label media as manipulative if it is shared in a deceptive manner and if the content is likely to impact public safety or cause serious harm.
Scavino’s video is edited to truncate Biden’s remarks: “We can only re-elect Donald Trump if in fact we get engaged in this circular firing squad here. It’s got to be a positive campaign.”
Labeling in Obscurity
Twitter developed its manipulative media policy in response to a survey of more 6,500 global users last fall, and following discussions with civil society and academic experts.
Nearly nine out of 10 individuals said placing warning labels next to content that was significantly altered would be acceptable, noted Yoel Roth, Twitter’s head of site integrity, and Ashita Achuthan, the platform’s group product manager.
“Overall, people recognize the threat that misleading altered media poses and want Twitter to do something about it,” they wrote in an online post.
How effective labeling manipulative media will be in reducing the flow of misleading information on the platform remains to be seen.
“It’s better than nothing, but it’s not a heck of a lot,” said John Carroll, a media analyst for WBUR in Boston.
“The first time I looked at a screenshot of it, I didn’t even see the warning,” he told TechNewsWorld.
The label doesn’t appear in the tweet detail, but is visible on timelines and in searches, Twitter spokesperson Katie Rosborough acknowledged.
“We’re working on a fix,” she told TechNewsWorld.
It might help to make the label larger.
“If you want an effective warning, you’d have a diagonal banner across the video that stays there for the entire length of the video,” Carroll recommended.
“That would be effective. What they’re doing now is not,” he said.
“I’d praise Twitter for trying to do something good, but I don’t know if it’s going to do any good given the times we live in,” said Dan Kennedy, an associate professor in the Northeastern University School of Journalism in Boston.
Taking down the content may be a better solution than labeling it, he continued.
“There’s a lot of research that suggests repeating or perpetuating a lie — or in this case, calling additional attention to the lie — only makes people believe it all the more, even though it’s labeled as ‘manipulated,'” he told TechNewsWorld.
In Twitter’s global survey, a majority of respondents (55 percent) supported removing or hiding manipulative media, while opponents raised concerns about free expression and censorship.
“Twitter is trying to point out misinformation while circumventing free speech issues,” he told TechNewsWorld.
“Some people might say this content should be banned altogether from the platform, which is a reasonable argument, but the approach Twitter is taking is interesting because it’s educating users about what content is real and what content is not,” Raynauld continued.
“It’s also adding a layer of meaning to the actions people are taking by posting this kind of content,” he added.
Tackling Deep Fakes
The core problem Twitter should be addressing is deep fakes, maintained John Sample, vice president of the Cato Institute, a public policy think tank in Washington, D.C.
“Manipulation of media online is developing in ways that make it difficult for people to determine the veracity of not only the message in a video, but the action, too,” he told TechNewsWorld.
“Twitter should be applying its manipulative media policy to deep fakes, instead of clips edited to misinform, but in a traditional way that can be corrected by counter-speech — people pointing out the manipulation on Twitter,” Sample said.
“They should focus on the central problem, which is deep fakes, because they can’t be corrected by counter-speech,” he said.
“If they go with these ‘edge cases,’ like the Biden video, I suspect they’re going to do too much and create problems for themselves with regulators and politicians,” Sample added.
The challenge for Twitter will be weeding out harmful manipulative media from manipulative media that’s not harmful, maintained Karen North, director of the Annenberg Program on Online Communities at the University of Southern California in Los Angeles.
“They will have a problem identifying problematic content,” she told TechNewsWorld.
“The same technology that goes into manipulated media goes into artistic media, so it’s going to be a challenge for them to identify manipulated media with sinister intent versus manipulated media with artistic intent,” North added.
Living Up to Responsibilities
Although Twitter tagged the Scavino video as manipulated content, Facebook let it run unchallenged.
Facebook hasn’t shown any inclination to label videos, NU’s Kennedy observed.
“Facebook remains far more powerful than any social media platform, and it’s not especially responsible in its behavior,” he added.
Social media platforms should do more to address the problem of distributing false content, media analyst Carroll maintained.
“They have the ability to at least police the content that’s brought to their attention and make a judgment about whether it should be there or not,” he said.
“This isn’t about misleading information. This is about content that can demonstrably be proven false,” continued Carroll.
“This is about the debate that Mark Zuckerberg and a whole lot of other people don’t want to have,” he added. “Facebook is a media company. It’s not a merely tech platform anymore, and it should be forced — either through public opinion or government regulation — to act like a media company.”