- Instagram has strengthened its default protections to reduce sexually suggestive and violent content in recommendations.
- Meta says users in India now report unwanted recommended content about half as often as they did a year ago.
- The changes come amid government scrutiny of how social media platforms amplify misleading and potentially inflammatory material.
Instagram has changed its recommendation systems in India to show users less sexually suggestive and violent content, as the government steps up scrutiny of how social media platforms decide what appears in people's feeds.
Meta, which owns Instagram, said on Monday (16) that it had strengthened the platform's default protections over the past year, using more advanced artificial intelligence (AI) models to identify content that does not meet its recommendation guidelines.
The changes apply primarily to content suggested from accounts users do not follow. These recommendations appear across Instagram's Explore, Reels and Feed sections, where the platform introduces users to posts beyond the accounts they have chosen to follow.
Meta said the improvements were guided by feedback from users in India about the material they did not want to see.
"We've improved our systems specifically for India so they catch more of what people here told us they didn't want to see," the company said in its announcement.
The company also reported that people in India now flag or report recommended content about half as often as they did a year ago. However, it did not disclose the underlying figures, the volume of content affected or the methodology used to calculate the reduction.
The announcement comes weeks after the Ministry of Electronics and Information Technology (MeitY) asked Meta to review its content-ranking and recommendation systems. The company has not directly linked its latest changes to the government's advisory, describing them instead as part of a year-long effort shaped by user feedback.
What has Instagram changed in India?
The update focuses on what Instagram recommends, rather than introducing a blanket ban on sexually suggestive or violent content.
The platform applies separate Recommendation Guidelines to content it suggests from accounts users do not follow. These guidelines are designed to prevent low-quality, objectionable or particularly sensitive material from being promoted, even when it does not violate Instagram's broader Community Standards.
Meta said it has improved its detection systems, including by using more capable AI models that assess content directly before it is recommended. This reduces the platform's reliance on users reporting or hiding posts before action is taken.
The company said its India-specific systems now identify more sexually suggestive material and content depicting violence, keeping a greater proportion of such posts out of recommendations.
This distinction matters because content that violates Community Standards is removed from Instagram, while material that does not meet recommendation guidelines may remain available on the platform without being actively promoted to wider audiences. Content that is unlawful in a particular country is restricted there, Meta said.
Users also have several tools to influence what appears in their feeds. Sensitive Content Control allows them to choose to see less sensitive material across Search, Reels, Explore, suggested accounts and in-feed recommendations.
Political Content Control lets users reduce political recommendations from accounts they do not follow. The Not Interested option allows users to signal that they do not want similar posts, while Hidden Words can filter selected words, phrases and emojis from comments and message requests.
Meta also said Teen Accounts, introduced in India last year, automatically place teenagers in a 13-plus content setting with age-appropriate protections.
Why did the government ask Meta to review its algorithms?
In an advisory dated October 10, MeitY asked Meta to examine how its ranking and recommendation systems distribute content relating to public events and developments.
The ministry was concerned that potentially misleading, manipulated or out-of-context material could receive disproportionate visibility or be repeatedly recommended to users.
It also asked the company to prevent its systems from exposing people predominantly to a single or limited set of narratives. The advisory called for stronger checks on content that could cause public confusion, incite violence or disturb public order.
The government's focus was therefore not limited to whether individual posts breached platform rules. It also wanted Meta to examine how its algorithms influence the reach and circulation of information.
The ministry asked the company to report the measures it had taken in response.
Meta's latest announcement does not explicitly refer to the advisory. Instead, the company said its changes had been developed over the past year in response to user feedback.
The company has previously adjusted Facebook's content-ranking systems in countries including Canada and Brazil to reduce the prominence of political content.
The latest Instagram changes add to the wider debate over how technology companies balance personalized recommendations, user safety and the spread of information.
For Instagram users in India, the immediate effect should be fewer recommendations featuring material the platform considers sensitive or objectionable. But Meta has not provided enough data to independently measure the scale of the change.
The broader question remains how much influence governments should have over the algorithms that shape what people see online, and how platforms can respond to concerns about harmful content without unnecessarily limiting access to legitimate information and discussion.

















Vijay Deverakonda's 'home' to Rashmika Mandanna in Japan post gets 'trying to prove you love her' reaction