|Founder and boss of Facebook, Mark Zuckerberg (Shutterstock).|
Facebook Mark Zuckerberg has posted a 5,500 word missive on the social network, in which he discussed many topics. One is the company's plan to overcome terrorist activities and "false news" by using AI and algorithms.
Global community of Facebook, with more than one billion people posting billions of messages every day, it becomes impossible for individuals to effectively supervise the overall network.
"The complexity of the problems that we have seen have exceeded the existing processes in the society. We are now exploring a system that can read text and view photos and videos to understand if something dangerous might happen," said Zuckerberg.
However, Zuckerberg admits that this system would require time to refine.
"It's still very early in development, but we have started to see it in some of the content. At present, we began to explore ways to use AI to distinguish between terrorism and propaganda news about the real terrorists," he continued.
This is an approach that can help turn against the flow and news sources "false". Even so, Zuckerberg also noted that humans can distinguish between a trusted content and unpleasant. Initially, there is a difficulty for the AI to understand it.
"It should be noted that great progress in AI are needed to understand the text, photos and video to assess whether they are hateful, graphic violence, explicit sexual content, and much more. We hope to start handling some cases in 2017, but several other cases can only be dealt with in the next few years, "he said.
So, even though for a long-term goal, the core of Facebook's defense will remain in the community moderation. [Tech Radar]