Facebook Doesn't Want to Censor Political Ads Over Accuracy, Executive Says
October 22 2019 - 1:30AM
Dow Jones News
By Georgia Wells
LAGUNA BEACH, CALIF.-- Facebook Inc. doesn't want to censor paid
content from politicians based on accuracy but it would draw the
line at political ads if they encourage violence, Chief Technology
Officer Mike Schroepfer said Monday.
Facebook wants to moderate content as little as possible
because, as a private company, it doesn't want to make decisions
about speech, Mr. Schroepfer said, echoing recent comments from
Facebook CEO Mark Zuckerberg. But an ad posted by a politician that
might inspire imminent harm to another person would be barred from
the platform, Mr. Schroepfer said at the WSJ Tech Live conference
here.
"It's pretty high up in the policy ranking of the things we
don't want on the platform," he said. "If this speech is going to
cause...harm, I think most people think that's a good place to
start intervening."
Facebook angered Democrats earlier this month with its refusal
to remove an ad from President Donald Trump's re-election campaign
that made an unsubstantiated claim about former Vice President Joe
Biden's role in the ouster of a Ukranian prosecutor.
Facebook denied the Biden campaign's request to remove the ad,
which the campaign said was false. That decision was made according
to a policy Facebook announced in September that it won't
fact-check speech or advertising by politicians.
Political advertising has caused a disproportionate headache for
Facebook compared with the amount of revenue it provides the
company. "We have had heated debates about whether political ads
are worth it," he said. "There's an argument that it's getting
abused too much, and there's an argument about access."
Elizabeth Warren in early October accused Facebook of
prioritizing profit over protecting democracy.
Mr. Schroepfer said he believes political advertising is good
because it could allow potential candidates who can't afford
broadcast ads to still promote themselves.
Facebook is still building out other aspects of its advertising
policy, including how the company should treat content manipulated
with artificial intelligence or other technology, known as
"deepfakes," Mr. Schroepfer said. "This is a part of our policy
that is still in discussion."
If a video has been edited it may be eligible for fact-checking,
and potentially labeled as misinformation. "There's not a policy
yet for how do you treat it just because it is an AI-created
thing," he said.
Twitter Inc. said earlier at the conference that it is planning
a new policy for deepfake content.
Meanwhile, Facebook is also working to better train its
algorithms to detect violent videos, Mr. Schroepfer said. The
company hopes to prevent a repeat of the March tragedy when a video
of the massacre in Christchurch, New Zealand remained on Facebook
for an hour before the company took it down.
"AI systems need examples to understand what they're looking
for. We literally didn't have a lot of first-person examples like
that," he said.
Since Christchurch, Facebook has partnered with the London
Metropolitan Police Service to get more data related to
first-person attacks. Facebook provided the police with cameras,
and the police are running training simulations of these types of
attacks.
"We're going to feed that straight into our systems to detect
this," he said.
Write to Georgia Wells at Georgia.Wells@wsj.com
(END) Dow Jones Newswires
October 22, 2019 01:15 ET (05:15 GMT)
Copyright (c) 2019 Dow Jones & Company, Inc.
Meta Platforms (NASDAQ:META)
Historical Stock Chart
From Aug 2024 to Sep 2024
Meta Platforms (NASDAQ:META)
Historical Stock Chart
From Sep 2023 to Sep 2024