Instagram now uses machine learning to detect intimidation within photos – TechCrunch



[ad_1]

Instagram and its users enjoy the ownership of the application by Facebook, which invests heavily in new technologies of artificial intelligence. Now that AI could help keep Instagram more tolerable for humans. Today, Instagram has announced a new set of anti-cyberbullying features. More importantly, he can now use machine learning to digitally scan photos posted on the app to detect bullying and send the message to the Instagram community reviewers for review. This means that stalkers will not be able to simply scribble threatening or defamatory notes and then post a photo of them to bypass Instagram's text filters for bullying.

In his first blog post addressing directly to Instagram users, Adam Mosseri, the new head of the division, writes: "There is no room for intimidation on Instagram. . . As the new Instagram Manager, I'm proud to reinforce our commitment to making Instagram a generous and safe community for all. The filter for photos and captions will be deployed in the coming weeks.

Instagram launched text filtering for bullying in May, but that could have pushed the trolls to attack people through images. Now, her bully classifier can identify harassment in photos, including insults to a person's character, appearance, well-being, or health. Instagram confirms that the image filter will work in streams and stories. "Although this update focuses only on the photos, we will work to add protections for the video, including IGTV, very soon," said a spokesman.

Instagram users will see the "Hide Offensive Comments" setting enabled by default in their settings. They can also choose to manually list the words they want to filter in their comments and choose to automatically filter the most frequently reported words. The text is in black and white, which allows Instagram to block keywords. With images, artificial intelligence will not be able to play the executioner; instead, she uses the filter to direct messages to the human moderators making the last call.

At the same time, Instagram is expanding its proactive anti-bullying filter in Stream comments, Explorer and Profile, to protect live broadcasts as well. He launches a "Kindness" camera effect in partnership with Maddie Ziegler, better known as the children's version of Sia in his "Chandelier" clip. The effect imbues your image with hearts and positive comments in different languages ​​while inviting you to tag a friend who is close to your heart. Users can view the camera effects tray if they follow Ziegler, or if they see a friend using it, they can try it themselves.

In order for Instagram to remain the teenagers' favorite application, we must not let this vulnerable community become a victim. There has been much talk that Facebook interferes with Instagram after the co-founders of the photo application quit. But the huge engineering organization of the parent company allows Instagram to realize economies of scale that unlock technologies such as this anti-bullying filter that an independent startup might not be able to develop.

[ad_2]
Source link