Meta has announced significant improvements to Instagram's content recommendation systems specifically for India, aiming to reduce unwanted content like sexually suggestive or violent material. The company states these changes have halved user reports on recommended content over the last year.
Meta has rolled out an update to its Instagram content recommendation systems, with specific enhancements targeting Indian users to reduce unwanted content. According to an announcement from Meta Newsroom, these changes are designed to improve people’s experience on Instagram by better filtering out objectionable or sensitive material.
Over the past year, Instagram has strengthened its default protections for recommended content in India. The company noted that its systems are now more effective at detecting and avoiding recommendations for content that Indian users have indicated they do not wish to see. This includes categories such as sexually suggestive content and content depicting violence. As a direct result of these improvements, Meta reports that users in India are now reporting recommended content about half as often as they were a year ago.
How Instagram's Recommendation Engine Works
Instagram's recommendation engine aims to help users discover new communities and content by providing personalized suggestions. If a user frequently interacts with accounts related to, say, restaurants or bookstores, the platform might recommend content about food, recipes, books, or reading. These recommendations appear in sections like Explore, Reels, and Feed, often featuring content from accounts a user doesn't already follow.
The platform applies specific guidelines to this recommended content, working to avoid low quality, objectionable, or particularly sensitive material. While content that violates Instagram's Community Standards is always removed, and unlawful content is restricted in specific countries, the recommendation guidelines serve as an additional layer of filtering.
Smarter AI and India-Specific Tuning
A key part of this update involves better detection capabilities. Instagram has improved its systems globally, incorporating more capable AI models that proactively assess content against its Recommendation Guidelines. This means the AI can identify and filter out unwanted content directly, rather than waiting for users to report or hide it.
Crucially, these systems have been specifically tuned for the Indian market. Meta stated it has listened to feedback from Indian users and adjusted its AI models to better identify and suppress the types of content people here found undesirable. This localized approach is what has led to the noticeable reduction in user reports.
User Controls for a Personalised Feed
Beyond these platform-wide enhancements, Instagram continues to offer several user controls to customise their content experience. These tools empower users to further adjust what they see and who can interact with them:
- Sensitive Content Control: Located in Settings and Activity, under Content Preferences, this allows users to choose to see less sensitive content across recommendations in Search, Reels, Explore, suggested accounts, and in-Feed. For India, where the most mature content is already not recommended, this control fine-tunes what appears within those existing boundaries.
- Political Content Control: Also found under Content Preferences, this option lets users reduce the amount of political content they see from accounts they don't follow.
- Not Interested: Tapping the three dots on any recommended post or reel allows users to signal that they don't want to see similar content in the future.
- Hidden Words: This feature filters out specific words, phrases, and emojis from comments and message requests, offering a layer of protection against unwanted interactions.
- Block: Users can block accounts to prevent them from finding their profile, seeing posts, or sending messages. This also extends to any new accounts the blocked individual might attempt to create.
Meta also highlighted existing protections for younger users in India. Last year, the platform introduced Teen Accounts, which automatically apply a 13+ content setting shaped by age-appropriate content guidelines for users aged 13 and above.
Meta affirmed its commitment to continuously improving these systems as user behaviour and technology evolve. The company will continue to consider feedback from people in India and regularly publishes data on its policy enforcement in its global Community Standards Enforcement Report and its monthly transparency report for India, as mandated by the IT Rules, 2021.