In Facebook actively fighting with an illegal content on the pages of his social network. In this struggle the workers of the company actively helps artificial intelligence. However, not all the tasks artificial intelligence to cope well enough, this was told by CEO of Facebook, mark Zuckerberg, communicating with analysts during the discussion of the quarterly performance of the company.
The involvement of artificial intelligence in the moderation of content on Facebook is growing. Still, one cannot completely rely on technology. Mark Zuckerberg believes that artificial intelligence is good. In particular, there are big problems with understanding human speech.
In Facebook very effectively removed publications that are associated with the promotion of terrorism. According to the CEO of the social network, is the merit of AI. In addition, the artificial intelligence is extremely effectively finds pictures of naked people. But the insults and angry publishing is the weak point of the technology.
“It is easier to teach artificial intelligence to find nipples on the images than to recognize abuse,” says mark Zuckerberg. Using the AI 99% of publications that promote terrorism, are removed without the participation of the company’s employees, but the achievement of effective search abuse can take years, says the founder of Facebook.