Photo and video sharing network Instagram has announced that it has trained an artificial intelligence (AI) system to detect offensive captions on a photo or video.
In a blog posted on Monday, Instagram announced the feature which will enable an AI system to warn users when their captions on a photo or video could be considered offensive.
The Facebook-owned company said the feature was developed to give users "a chance to pause and reconsider their words". Instagram said the system will be implemented immediately in some countries.
The AI system is designed to help address the problem of online bullying, which has become a major concern among online platforms such as Instagram, YouTube, and Facebook. A cyber-bullying study in July 2017 ranked Instagram as the worst online platform.
Once implemented, the tool will send a prompt to a user typing an offensive caption on Instagram informing them that it is similar to others reported for bullying. The users will then be allowed to edit their caption before it is published.
Instagram wrote on the blog post: "In addition to limiting the reach of bullying, this warning helps educate people on what we don't allow on Instagram and when an account may be at risk of breaking our rules."
Earlier this year, the company launched a similar feature that notified users when their comments on other people's Instagram posts could be considered offensive.
"Results have been promising and we've found that these types of nudges can encourage people to reconsider their words when given a chance," Instagram added.
Chris Stokel-Walker, internet culture writer and author of the book YouTubers, said the feature was part of a broader move by Instagram to be more aware of the well-being of its users.
Earlier this month, Instagram announced that it will require new users to provide their date of birth when creating a new account on its platform. According to Reuters, this move will help the company avoid targeting ads at children for age-restricted products.
