Facebook knows how to intercept more and more online hate speech with special software.
The American company filters banned content on the platform with artificial intelligence and say it is making considerable progress.
In the past quarter, the software detected 97 percent of hate speech before users even saw the messages, Facebook said.
The previous quarter it was 94 percent, a year earlier just 80 percent – the rest of the tracking work was done by inspectors, who manually removed the hate speech.
The social network is confident that the software can detect incitement to hatred in texts and in sound and videos on Facebook.
“We understand audio and video, we understand the content,” said the technical manager of the social network.