06:51 GMT +304 December 2016
Live
    Facebook

    Facebook Hopes Artificial Intelligence Can Stop Offensive Live Streams

    Pixabay
    Tech
    Get short URL
    424135

    Facebook has been on a mission to rid itself of offensive content, and is now working on using proprietary artificial intelligence to flag live video streams.

    The company has traditionally relied on users to report offensive posts, but is hoping to automate its censorship. 

    Joaquin Candela, the company’s director of applied machine learning, said during a news conference that the company has been increasingly using artificial intelligence to moderate offensive content. The algorithm that they are using, according to Candela, currently “detects nudity, violence, or any of the things that are not according to our policies.”

    In June, Facebook began using automation to flag and remove uploaded extremist video content, but there are currently no algorithms to detect the content in live video feeds. Social media platforms have been under tremendous pressure from governments to quickly remove violent propaganda from organizations such as Daesh.

    The platform also uses automation for user reports, recognizing duplicate reports and routing flagged content to reviewers who have expertise in the content’s subject matter, Reuters reports.

    Candela stated that Facebook has identified two challenges in automating live video content removal.

    “One, your computer vision algorithm has to be fast, and I think we can push there, and the other one is you need to prioritize things in the right way so that a human looks at it, an expert who understands our policies, and takes it down,” he said.

    Facebook is also planning to start using automation to identify “fake news,” after Democratic outcry claiming that alternative media stories containing false information helped Donald Trump win the election.

    Platforms that use automation to remove posts are generally quiet about their methods.

    “There's no upside in these companies talking about it,” Matthew Prince, chief executive of content distribution company CloudFlare told Reuters in June. “Why would they brag about censorship?” 

    Related:

    From Fake News to Censored Posts: Facebook Backlash Over China Suppression Tool
    Facebook Develops Censorship Software in Bid to Re-Enter Chinese Market
    Facebook Is Going to Fight Misinformation – But Is ‘Fake News’ the Problem?
    Facebook CEO Mark Zuckerberg Talks Tech, AI and Health Research at APEC Summit
    Facebook’s Zuckerberg Headed for APEC Summit in Lima
    Tags:
    Live Streams, Censorship, Artificial Intelligence, Facebook, Joaquin Candela
    Community standardsDiscussion
    Comment via FacebookComment via Sputnik
    • Сomment

    All comments

    • Antares Prime
      Facebook's entire business model is built on intruding deep into people's lives, making people transparent instead of making itself transparent, and then selling that information to other companies. It has tried to replace friendship with a "friend" button, and people buy into it. People are complicit in turning their own personal lives into a commodity.

      If people agree to fall under that kind of control mechanism, why will they care if the great technological demigod decides what to post and what not to post? It is only censorship if it comes from the government.

      For anyone who doesn't think Facebook is a force for good in the world, the solution is simple. Stop feeding it.
    • Mraz
      Yep, time to ditch FB.
    • michael
      well, it doesn't seem as if there is too much human 'intelligence' there at fb.
    Show new comments (0)

    Top stories