Bloomberg Anywhere Login

Bloomberg

Connecting decision makers to a dynamic network of information, people and ideas, Bloomberg quickly and accurately delivers business and financial information, news and insight around the world.

Company

Financial Products

Enterprise Products

Media

Customer Support

  • Americas

    +1 212 318 2000

  • Europe, Middle East, & Africa

    +44 20 7330 7500

  • Asia Pacific

    +65 6212 1000

Communications

Industry Products

Media Services

Follow Us

Facebook

Will Facebook's Robot Censors Know Porn When They See It?


Will Facebook's Robot Censors Know Porn When They See It?

Photograph by George Logan

The old trope about pornography is that you’ll know it when you see it. Facebook (FB) is betting that not only will it know it, but that the social network will be able to train its computers to do the same.

On Friday the company announced a new advertising policy for potentially controversial pages, saying it will ban ads from pages or groups containing violent, graphic, or sexual content. The company is responding to a flap last month in which brands began pulling advertising when they found that it was showing up next to pages with crude, sexist content. In an unrelated move, Google (GOOG) said it will soon prohibit users of its Blogger service from carrying ads for adult businesses.

These issues have become regular irritants for Internet companies whose sites are full of user-generated content. Playing referee is not a comfortable role. Not only do companies face the choice of appearing to be either prudish or insufficiently sensitive to offended parties, it’s simply difficult to monitor millions of pieces of content to determine whether each one violates some sensibility or another. Facebook’s judgement has been called into question in the past, with the company deleting images of gay men kissing while letting pages full of rape jokes remain on the site. And the job of reviewing potentially offensive content on social media is a low-paid and dehumanizing job.

This time around, Facebook is hoping to use a short period of human-based review before passing the job to computers, the company explained in describing its new policies:

In order to be thorough, this review process will be manual at first, but in the coming weeks we will build a more scalable, automated way to prevent and/or remove ads appearing next to controversial content. All of this will improve detection of what qualifies as questionable content, which means we’ll do a better job making sure advertising messages appear next to brand-appropriate Pages and Groups.

The idea of computer programs that can identify pornographic photographs through algorithmic analysis that identifies, say, a lot of uncovered skin, is not new. But Facebook’s challenge goes beyond that: It must determine which content is inappropriate in which context. Even people have had trouble making these calls. It will be interesting to see how well Facebook’s computers can do it.

Brustein is a writer for Businessweek.com in New York.

LIMITED-TIME OFFER SUBSCRIBE NOW
 
blog comments powered by Disqus