Facebook-owned Instagram is ramping up efforts to address the issue of offensive comments by activating new artificial intelligence-powered features.

The engine enabling these features, named DeepText, is an AI engine that is capable of analyzing both the content and the context of comments to determine if they are offensive or spam. The system was originally built for Facebook, so should users expect the same comments moderation by an AI to soon arrive on the social network?

Keeping Instagram Safe

In a blog post, Instagram CEO and cofounder Kevin Systrom introduced the two new tools that will "help keep Instagram a safe place for self-expression."

The first tool is a filter that will block offensive comments that other users leave on posts and in live video. All other comments will appear as they normally do, and users will still be able to report comments if there are any offensive ones that slip through.

The feature can be deactivated by going into Instagram's settings menu from the user profile and tapping on Comments, though it is unclear why any user would want to do this. The filter against offensive comments will first be rolled out for English, with support for other languages to come over time.

The second tool is another filter that will block obvious spam comments in posts and live videos. Unlike the filter against offensive comments, the spam filter will initially already support several languages, including English, Arabic, Chinese, French, German, Japanese, Portuguese, Russian, and Spanish.

According to Systrom, the Instagram team has been training the system behind the tools for some time and that the tools will continue to improve over time. Instagram believes using machine learning in building tools that protect self-expression will lead to inclusive and kinder communities.

"Our work is far from finished and perfect, but I hope we're helping you feel safer and more welcome on Instagram," Systrom said to end the blog post.

DeepText Enables AI Features On Instagram

Wired published an in-depth report on the capabilities of DeepText and the new Instagram tools, explaining that artificial intelligence is necessary so that machines can understand the meaning of words and sentences as well as humans.

DeepText was initially developed as an in-house tool to help Facebook engineers quickly sort through huge walls of text so that they can build products that can help users. However, when Facebook acquired Instagram in 2012, the picture-sharing social network's executives saw the potential of DeepText in fighting against offensive comments and spam.

Training DeepText for such a purpose was a long process, though, beginning with a team of humans having to sort all the comments on Instagram as either spam or not spam. From eliminating spam comments, the system's purpose on the social network eventually expanded to hunting down offensive comments.

Facebook Offensive Comments: Next Target?

With Facebook owning Instagram, it could be assumed that the DeepText-powered filters will eventually arrive on the social network. Facebook even recently mentioned that it was attempting to use AI to fight against hate speech.

However, a spokesperson for Facebook told Variety that the two platforms are different, which means that different approaches will be required to reach the goal of creating safe communities. However, the spokesperson added that the new tools on Instagram will help both platforms learn about the capabilities of DeepText, which means that the engine will likely find its way into Facebook in one form or another in the future.

ⓒ 2024 TECHTIMES.com All rights reserved. Do not reproduce without permission.
Join the Discussion