FACE­BOOK RE­MOVES SPAM, BUT HATE SPEECH HARDER

USA TODAY International Edition - - MONEY -

MONEYLINE

Face­book for the first time pub­lished in­for­ma­tion on how much and what kind of con­tent it re­moves. The 86page re­port showed the gi­ant so­cial net­work is mak­ing progress us­ing ar­ti­fi­cial in­tel­li­gence to de­tect and re­move nu­dity, spam and ter­ror­ism, but it isn’t yet able to com­bat all of­fen­sive con­tent, es­pe­cially hate speech. Face­book says it re­moved 865.8 mil­lion posts in the first quar­ter of 2018, most of which were spam (837 mil­lion). Ar­ti­fi­cial in­tel­li­gence found 99.5% of ter­ror­ist con­tent and re­moved about 1.9 mil­lion items. But of the 2.5 mil­lion items of hate speech Face­book re­moved in the first quar­ter, only 38% were de­tected by ar­ti­fi­cial in­tel­li­gence.

MATT ROURKE/AP FILE PHOTO

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.