Meta has updated Instagram’s recommendation algorithms in India so that sexually suggestive and violent content from accounts you don’t follow isn’t visible.
The update is about improving the content experience on Instagram Reels, Explore and Feed using AI (artificial intelligence) systems to identify and limit potentially sensitive material.
These changes are part of Meta’s broader effort to make Instagram safer users often see recommended posts from unknown accounts changing how content is selected and displayed helps reduce unwanted exposure to posts that aren’t appropriate for all audiences.
Meta has used AI technology to improve its content recommendation systems information about the update. Those systems have also been tuned with local feedback from Indian users so that the platform can better address concerns about what type of content is shown in recommendations.
The impact of these changes is seen in user reports. Complaints about sexually suggestive and violent content have reportedly dropped half as much this year as last year (i.e., complaints that such material was appearing in users’ recommendations).
The new strategy only applies to posts that Instagram recommends not the ones users choose to access, so some posts may no longer be visible in recommendations but won’t be removed from the platform.
They can still see content from accounts they follow. They can search for particular accounts or posts if they want to find them. That’s what allows Instagram to limit unsolicited recommendations but not completely block access to content that people actively seek.
However the changes have also raised questions about how Instagram classifies sensitive material. Some users have reportedly seen restrictions on neutral news posts. This has highlighted the difficulty of using automated systems to tell the difference between truly harmful content and content shared for informational or journalistic purposes.
News stories, public discussions and other informational posts may have images or references to violence but are not promoting it. If recommendation systems classify that material too broadly legitimate content could also be lost.
Meta has made it clear that people have control over settings that change how they view content. Sensitive Content Control is an example of this users can opt to see less or more potentially sensitive material in recommended content.
These settings give people more options if they want to limit their browsing experience or see a wider range of content.
The latest Instagram update shows how much Meta is relying on AI to manage recommendations and respond to user feedback. Social media platforms are still grappling with issues around harmful content but balancing safety for users with access to news and information remains an important challenge.
For Instagram users in India the changes could mean fewer unwanted sensitive recommendations but still be able to find content and accounts they want to follow.
Comments