The number of fake accounts is striking.
Facebook in the first quarter of 2018 removed 583 million bogus accounts, or 3-4% of the monthly audience. Typically, these accounts deleted within a few minutes after their creation.
Facebook removed 583 million dummy accounts in the first quarter. It is reported by the Chronicle.info with reference to eizvestia.
On Tuesday, the largest social network in the world for the first time published a report on the work of the artificial intelligence (AI), which should protect users from the stream of false information, pornography and messages of hate. The document is 86 pages.
According to Facebook, in January-March was removed 865,8 million posts that do not meet the requirements. Most often they contain spam.
Over the last 18 months Facebook has significantly increased the measures aimed at identifying inappropriate content and protect users, said Vice-President of product management, guy Rosen.
In January-March AI, estimated Facebok, showed 99.5% of 1.9 million attempts to post messages containing calls for terrorism. In addition, the system identified 95.8 per cent of the 21 million posts with nudity and 85.6% of the 3.4 million scenes of violence.
However, messages inciting hatred, the robot can distinguish much more difficult — a computer’s intelligence does not understand the cultural and contextual features, so in this category, Facebook is largely dependent on staff moderators.
Only 38% of the deleted posts with hate speech and xenophobia was discovered AI.
The report did not contain quantitative indicators, characterizing the struggle of the social network with fake news.