Yahoo gets ready to tackle online hate speech

People often miss the thin line dividing freedom of expression from "hate speech" as many internet users believe it can boost their social standing since they walk in digital steps - Facebook, Twitter, and Snapchat. Hence, the research team at Yahoo has come up with an "automated abuse filter" which is created from abusive online comments to detect and control online abuse.

Applying a first-of-its-kind data set created from offensive article comments flagged by Yahoo editors, the research team was able to develop an algorithm that can recognize offensive words on their own. Using a technique called "word embedding," this algorithm will process words as vectors rather than simply positive or negative.

The current abuse filters rely on a combination of blacklisted terms, common expressions, and syntax clues to catch hate speech online, however, the Yahoo team has gone a step further and applied machine learning to their massive closet of flagged comments.

According to their findings, the system was able to correctly identify abusive language from the same data set about 90 percent of the time.

Though the results are impressive, Alex Krasodomski-Jones, an online abuse researcher with the UK's Centre for Analysis of Social Media claims that the ever-changing nature of hate speech means no system - not even a human one - will ever truly be able to know what offensive 100 percent of the time is.

He told Technology Review, "Given 10 tweets, a group of humans will rarely agree on which ones should be classed as abusive, so you can imagine how difficult it would be for a computer."