By Yoree Koh And Reed Albergotti
Twitter Inc. this week appeared to be taking its most aggressive
steps ever to scrub sensitive content from its site as it scrambled
to remove grisly images showing the beheading of a U.S.
journalist.
In reality, the company is maintaining the same reactive
approach that has allowed violent and pornographic content to
proliferate on the social-media service.
Like its social-media brethren, Twitter tries to walk a fine
line between a dedication to free expression, an aversion to being
held legally responsible for the actions of its users, and the
reality that people will use such platforms to share highly
offensive material--a particular concern as it seeks to expand its
advertising business.
Twitter's broad free-speech stance came into focus Tuesday, when
the company said immediate family members could email a special
address to request the removal of images and videos of deceased
individuals.
Twitter said it made the decision "in order to respect the
wishes of loved ones" after doctored photos of Robin Williams's
death prompted his daughter Zelda to leave the service. The move
appeared to hint that Twitter, in a special case, would actively
try to stamp out sensitive content, rather than wait for users to
flag each offending tweet as it has been doing.
Tuesday afternoon, at the request of James Foley's family,
Twitter began taking down gruesome images and video depicting the
Islamic State's beheading of the American journalist. Twitter CEO
Dick Costolo tweeted to his 1.2 million followers, "We have been
and are actively suspending accounts as we discover them related to
this graphic imagery. Thank you."
More than 24 hours later many of the images remain in
circulation, angering some users. At first glance, it might appear
that Twitter's staff is playing a game of cat and mouse as it tries
to delete controversial content amid the service's half-a-billion
tweets a day.
But Twitter is actually sticking to its basic principles. Even
after a family request, it's the company's policy not to hunt for
content. Instead, it relies on users to flag a tweet as
inappropriate, at which point, if the tweet violates its rule,
Twitter will disable the unique Web address associated with the
image. But users can easily upload the image using a different
account and a new address, as has happened with images of Mr.
Foley's execution. Other images remain if they're not flagged.
It is unclear how many and at what pace users flagged the grisly
images of Mr. Foley's murder to Twitter and how quickly the service
was able to respond. Twitter declined to say.
Had Twitter not announced the family policy hours earlier, it
might not have taken any action. That's because Twitter doesn't ban
violent content such as beheadings on its own, just violent images
it deems to be direct threats to other users.
For example, a video showing St. Louis police officers fatally
shooting a 25-year-old man outside a convenience store on Tuesday
is still up on Twitter. Twitter declined to say whether it has
received requests from Kajieme Powell's family to remove the
video.
Twitter has removed images of beheadings in the past, but in
instances where the photos of severed heads were directed at users
as threats. In a statement, Twitter said: "We evaluate and refine
our policies based on input from users, while working with outside
organizations to ensure that we have industry best practices in
place."
As the self-described "free speech wing of the free speech
party," Twitter has been home to the profound and profane. The
company allows pornographic images in tweets if they aren't used to
harass other users. Twitter introduced a measure last year, marking
some "sensitive" images with a warning.
But as Twitter's user base has swelled to 271 million monthly
active users, the company may be reaching a point where it can no
longer remain the Wild West of social media without irking too many
members. Responding to the viral nature of the images of Mr.
Foley's death, the hashtag #ISISMediaBlackout, which urged users to
stop sharing the images, trended on Twitter earlier this week.
Twitter's policy concerning violence is more liberal than Google
Inc.'s YouTube, which itself has been trying to suspend the
accounts of those posting video of Mr. Foley's beheading. YouTube
prohibits "gratuitous violence," including hate speech and
incitement to commit violent acts, but it does allow violence if it
isn't designed simply to shock or be disrespectful. Like Twitter,
YouTube takes action only when users flag the content.
YouTube also bans designated foreign terrorist organizations
from having registered accounts. A Twitter spokesman said a group's
status as a terrorist group is "one of several factors" it
considers when deciding whether to suspend an account.
Facebook has had its own battles with violent content. In
October, Facebook faced international criticism for allowing its
users to share several videos depicting beheadings. Facebook had
initially allowed the videos because they were being used to
condemn violence, but it removed them following the public
outcry.
After the controversy over the beheadings, Facebook amended its
policies, adding more requirements for users who want to share
graphic content. The photos or videos should include safeguards
that prevent minors from viewing them, warnings about the nature of
the content and edits to remove excessively graphic content.
Facebook's efforts to block images and videos of Mr. Foley's
execution have largely been successful, according to a company
spokeswoman. Like Twitter, Facebook only blocks content reported by
users that violates its rules, but it also uses technology to
identify offending videos or photos across the site. Facebook also
removes content tied to individuals or groups deemed to promote
terrorism.
Twitter declines to comment on how it decides which accounts to
suspend and which content to remove. Every report by a user is
reviewed by a member of Twitter's Trust and Safety team. Depending
on the case, Twitter's lawyers may get involved. Twitter tells the
user who flagged the content what it ultimately has decided to do
but doesn't disclose the reasons for its decision.
"In this case, I don't think Twitter has done enough to be
consistent and transparent," said Jillian York, the director for
international freedom of expression at the Electronic Frontier
Foundation. "They need to be more transparent about what goes
against their terms. It's unclear what they're taking that content
down under."
Still, others are less critical of Twitter's overall methods to
manage its content. "Active policing of tweets is tricky, and I
think Twitter has made the right decision in saying that it will
not do that," said Jonathan Zittrain, a professor at Harvard Law
School. But he recognizes the limitations: "The fact is that unless
Twitter's membership were to be circumscribed, shutting down an
abusive user account is a necessarily limited remedy: An aggressor
can sign up again and take up where he or she left off."
Mr. Zittrain said "the best approach [for Twitter] may
ultimately be to sift garden-variety rudeness from genuinely
threatening behavior."
But to do this, Twitter would require more tools. It uses
PhotoDNA, a Microsoft image-tagging technology, to limit the spread
of child-sexual-exploitation images, as does Facebook. Besides
that, it doesn't actively monitor user content.
It could eventually get some help from Madbits, a New York
startup it acquired last month. Madbits's image-search technology
is designed to detect context in photos, and Twitter may be able to
use the Madbits technology to track sensitive images.
Write to Yoree Koh at yoree.koh@wsj.com and Reed Albergotti at
reed.albergotti@wsj.com
Subscribe to WSJ: http://online.wsj.com?mod=djnwires