#heartsnothate

Instagram’s New Machine Learning Tool Deserves a Double Tap

Social media has created arguably the most connected generation in history. Twitter (2006), Facebook (2006) and Snapchat (2011) have evolved over the years to create platforms that showcase instant, digestible content for their audiences.

Launched in 2010, Instagram soon rose to the top of the crowded social media vertical. It revolutionized the way content was shared on social media, in that it relied on the visual media more than traditional language based content. Sure, the caption and hashtaggery is still valuable, but most of the time I find myself double tapping the photo for its visual appeal before reading the caption.

Instagram found a way to cater, very effectively, to the visually driven consumer.

With this emphasis on attractive, curated visuals come the pitfalls of a culture already obsessed with “image”. The most connected generation scrolls through thousands of posts daily that flood their minds with the beauty standards from across the globe. It is natural to start comparing your content, your image, and your brand to all of the other social influencers out there…and it’s very easy for other’s to do so as well.

Bullying on social media has become a cultural epidemic. Trolls, bots, and hate groups have access to a much larger pool of potential victims thanks to the internet. Almost 21% of all teens surveyed in a national study admitted that they experience some form of bullying. Cyberbullying opened up a world where the bullying doesn’t end when you leave the school building, but can continue 24/7.

Though technology has managed to worsen the bullying epidemic, it can also be leveraged in the fight against the issue. Instagram is piloting a whole new set of anti-bullying features, powered by AI and machine learning.

The app can now use machine learning algorithms to scan photos posted on the app, detect signs of abuse or bullying, flag the post and send it to human moderators for review. This closes a previous loophole to the app’s ant-bullying/defamation policy that only screened the text in the captions for signs of policy violation, not the photos themselves which can contain text. This feature will work both on traditional Instagram posts and on their popular Stories feature.

In addition, the app now allows users to enable a “Hide Offensive Comments” setting, that automatically filters comments containing Instagram’s pre-made list of offensive terminology. Users also have the option to manually add words that they, personally, want filtered out of their comments.

Hopefully, this solution contributes to a significant decrease in the amount of vitriol spewed across social platforms, so we can get back to the days of kitten videos, celebratory milestones, and healthy cultural exchange.