Advertisement
X

Instagram's Content Reset: Why Meta Is Changing What Users See In India And How Its New Rules Work

In a statement on its official website, Meta said it had strengthened its default protections using improved artificial intelligence (AI) models to identify content that does not meet its recommendation guidelines before users report it

Meta Changes Instagram's Content Recommendation System
Summary
  • Meta has strengthened Instagram's AI-based recommendation filters in India to reduce the visibility of sexually suggestive and violent content, particularly from accounts users do not follow.

  • The changes distinguish between content that violates Instagram's Community Standards and material that remains accessible but is not eligible for wider recommendation.

  • The announcement follows a MeitY advisory asking Meta to review its recommendation algorithms over the amplification of misleading and out-of-context content, raising questions about moderation, transparency and censorship.

Meta has announced changes to Instagram's content recommendation systems in India, tightening the way the platform identifies and distributes sexually suggestive and violent content. The changes, announced on November 16, aim to reduce the circulation of material users may find objectionable, particularly posts recommended from accounts they do not follow.

In a statement on its official website, Meta said it had strengthened its default protections using improved artificial intelligence (AI) models to identify content that does not meet its recommendation guidelines before users report it.

The announcement comes amid increasing scrutiny of social media platforms and their algorithms. In October, the Ministry of Electronics and Information Technology (MeitY) asked Meta to review its recommendation and ranking systems to prevent misleading, manipulated or out-of-context material from receiving disproportionate visibility.

The latest changes have renewed questions about how much control platforms should exercise over content distribution and where the distinction lies between moderation and algorithmic censorship.

What Are The Latest Changes Announced By Meta?

According to the company, its systems now use improved AI models to assess content proactively rather than relying primarily on user reports. Meta also said it had adapted its systems to identify categories of content that users in India wanted to see less of.

The company claimed that users in India were reporting recommended content approximately half as often as they were a year ago. However, Meta did not disclose the underlying figures or methodology.

Importantly, the changes do not amount to a blanket prohibition on such content. They primarily concern what Instagram actively recommends, particularly material originating from accounts users do not follow.

Instagram operates two distinct layers of content regulation. Its Community Standards determine what is allowed to remain on the platform, while its Recommendations Guidelines determine what can be promoted to a wider audience. Consequently, a post may remain accessible on an account without being recommended to non-followers.

Advertisement

How Does Instagram Decide What Content Users See?

Instagram's recommendation system helps users discover content beyond the accounts they follow. It analyses their interests and interactions to identify posts, videos and accounts they are likely to engage with.

These recommendations appear across Explore, Reels and suggested posts in the main Feed. Instagram's ranking systems assess both the content and the likelihood of a user finding it relevant.

In its November 16 announcement, Meta explained that recommendations are subject to separate guidelines intended to prevent objectionable or particularly sensitive material from being actively promoted.

Thus, sexually suggestive content or violent imagery may not necessarily violate Instagram's Community Standards but can still fall outside the criteria for recommendation.

Meta's latest changes seek to strengthen this filtering through AI-based content identification. However, automated systems can make mistakes, particularly when interpreting satire, news footage and political commentary. The company has not disclosed the extent to which human moderators review content flagged by these systems.

Advertisement

What Controls Do Instagram Users Have?

Meta provides users with several controls to customise the content they encounter.

The Sensitive Content Control allows users to reduce sensitive material appearing in recommendations across Search, Explore, Reels and the main Feed. Meta said the most mature content was already excluded from recommendations in India. Users can access this option under Settings and Activity, followed by Content Preferences.

Instagram also offers Political Content Control, which allows users to reduce political content recommended from accounts they do not follow.

The 'Not Interested' option lets users indicate that they do not want similar recommendations by tapping the three-dot menu on a post or Reel. Other tools include Hidden Words, which filters selected words and phrases from comments and message requests, and the Block feature, which prevents particular accounts from interacting with a user.

Why Is The Centre Scrutinising Meta's Recommendation Algorithms?

Meta's announcement comes amid increasing government scrutiny of how social media algorithms distribute information in India.

Advertisement

On October 10, 2026, MeitY issued an advisory asking Meta to review its recommendation and ranking systems on Instagram and Facebook. The ministry expressed concerns about misleading, manipulated or out-of-context content receiving disproportionate amplification.

The ministry also called for greater diversity in recommended information and flagged old or selectively edited visuals presented as current events, misleading claims about emergencies and law-enforcement action, and material depicting law-and-order situations without adequate context.

The advisory came amid political tensions and protests in Delhi over the Special Intensive Revision (SIR) of electoral rolls, although the ministry did not identify a particular incident as the reason for its intervention.

Meta has not explicitly linked its November announcement to the government's advisory. Reuters, in a November 17 report, said the company had not clarified whether the India-specific changes were a response to the government's intervention.

Advertisement

Does Restricting Recommendations Amount To Censorship?

The distinction between removing content and restricting its recommendation is central to understanding Meta's latest changes.

Content removal prevents material from remaining accessible, usually because it violates the company's policies or applicable law. Recommendation restrictions, on the other hand, limit distribution without necessarily removing the content.

For instance, a video depicting violence may remain accessible on the account that uploaded it, while Instagram may decide against recommending it through Explore or Reels.

The Centre's October advisory has also raised concerns about government intervention in algorithmic distribution. Speaking to The Indian Express, Apar Gupta, founder of the Internet Freedom Foundation, questioned the legal basis of the intervention and cautioned against arbitrary restrictions on lawful expression.

Instagram's recommendation systems are proprietary, leaving users with limited information about why particular posts are excluded. The increasing use of AI adds another layer of complexity, particularly when distinguishing sensitive material from news reporting and legitimate public-interest content.

Meta's latest changes underline the growing influence of recommendation algorithms in determining what information, opinions and images reach social media audiences. The challenge will be to make the platform safer without arbitrarily restricting legitimate expression and access to information.

Published At:
US