This Chrome extension from Alphabet hides toxic web comments


We love trying neat online tools. Last week, we told you about an AI-powered B&W photo colouriser. Now, we have a cool extension for you to try.

Alphabet’s Jigsaw division has released an experimental Chrome extension that’s designed to rid your web-viewing experience of toxic comments. Called Tune, it works across the most popular sites you’d typically find horrible comments, like YouTube, Twitter, Facebook, Reddit, and comment platform Disqus. Just install it from here, sign into your Google account, and choose the sites you want to filter.

The default filter, dubbed Toxicity, gets rid of all comments “likely to make people leave a discussion”. You can also switch over to a mix of filters, where you can toggle on/off whether Tune should get rid of “attack on identity”, “insult”, “profanity”, “threat”, or “sexually explicit” related-comments. There’s also a “Zen mode” that turns off comments and “volume levels” ranging from “quiet” to “blaring”.

These levels let you adjust the level of toxicity you want to see across the internet. The goal is not to create a family-friendly web (though that is an added perk if you want to keep your kids safe online), but rather, to help you better “focus on what matters”.

It’s an open-source extension powered by machine learning. And it’s built on Perspective, an API created in 2017 by Jigsaw and Google’s Counter Abuse Technology team. It’s currently in use by The New York Times and The Guardian to moderate comments and improve online discussions. If you try it, you’ll notice filtered comments appear as a dot, which can be clicked on to show the hidden comment.

However, Jigsaw warned that Tune is an experiment that doesn’t always accurately filter comments – hence it’s “experimental” label. For more about how the extension works, check out Jigsaw’s work and explainer on Github and Medium.





Source link

LEAVE A REPLY

Please enter your comment!
Please enter your name here