Google’s think-tank Jigsaw is working on an open-source AI to combat trolls

Alphabet, the parent company of everything Google, is working on an Artificial Intelligence based software to combat trolls on the internet. The software based on machine learning has been designed and developed by the company’s subsidiary called Jigsaw, previously known as Google Ideas.

Called Conversation AI, the software works by identifying abusive language, threats and harassment on the internet. Researchers at Jigsaw are attempting to understand “the nuances and context of abusive language at scale.” A recent story by Wired describes how Conversation AI was trained using 17 million comments left under New York Times’ stories and 13,000 discussions on Wikipedia pages. The report goes on to state that the AI can spot abuse with 92% accuracy and a “10% false-positive rate” compared to humans.

Here’s what Wired’s Andy Greenberg had to say about Conversation AI in his piece:

“I suggest the first ambiguously abusive and misogynist phrase that comes to mind: "What’s up, bitch?" Adams types in the sentence and clicks Score. Conversation AI instantly rates it a 63 out of 100 on the attack scale. Then, for contrast, Adams shows me the results of a more clearly vicious phrase: "You are such a bitch." It rates a 96. In fact, Conversation AI’s algorithm goes on to make impressively subtle distinctions. Pluralizing my trashy greeting to "What’s up bitches?" drops the attack score to 45. Add a smiling emoji and it falls to 39.”

Greenberg goes on to report that the Times is looking to use Conversation AI to filter abusive remarks from its comments section and Wikimedia is also considering adopting the new AI troll tagger.

Follow Us

Adamya Sharma

Managing editor, Digit.in - News Junkie, Movie Buff, Tech Whizz! View Full Profile