Centre tells Meta to overhaul content promotion system on Facebook, Instagram

India’s Ministry of Electronics and Information Technology has directed Meta to change how Facebook and Instagram recommend content. The central aim is to prevent users from receiving disproportionate exposure to one narrative or a limited group of narratives. The ministry also wants more diversity in relevant and credible information.

The advisory requires Meta to review recommendation and ranking systems. Misleading, manipulated, selectively edited, or out-of-context material should not be disproportionately amplified or repeatedly recommended. Meta must also strengthen rapid identification, review, and escalation of content that could cause confusion, apprehension, violence, or public-order problems.

The advisory was dated October 10 and came as protests continued in Delhi over electoral-roll revision. It does not identify a specific triggering incident. Meta must act consistently with the Information Technology Act, 2000, the IT Rules, 2021, and other applicable laws while cooperating with authorised government agencies.

A recommendation algorithm is a set of computer instructions that selects and ranks content for each user. Instead of showing every available post equally, it predicts what may be relevant or engaging. This shapes feeds, suggested videos, account recommendations, and other discovery features.

The system can examine signals such as what someone watched, liked, shared, searched for, or skipped. It may also consider the content’s topic, how recently it appeared, and how many other users interacted with it. It then assigns an estimated relevance or engagement value and places some items more prominently than others.

The article does not describe Meta’s exact formula. Generally, recommendation systems matter because they can give selected content far more visibility than ordinary posting alone. That is why MeitY asked Meta to check whether recommendation and ranking systems repeatedly amplify misleading, manipulated, or out-of-context material instead of providing diverse, credible information.

MeitY warned against disproportionate amplification of material that can distort people’s understanding of current events. The advisory particularly focused on content that presents old or selectively edited visuals as if they were happening now. It also covered misleading claims about emergencies, restrictions, or police action.

Another concern involved audio-visual material showing law-and-order situations without adequate context. The ministry also flagged content likely to incite violence and misleading material receiving disproportionate amplification. Under the advisory, such material should not be repeatedly recommended simply because recommendation systems give it strong reach.

This matters most during fast-moving public events, when users may have little time to verify what they see. The advisory asked Meta to identify and address prohibited content, and to strengthen rapid review and escalation when recommended content could cause widespread confusion, apprehension, violence, or public-order disturbance.

Ordinary person-to-person sharing usually moves through direct contacts, private groups, or a person’s existing followers. A recommendation system can go beyond those circles by placing a post, video, or account in many users’ feeds or suggestions. Its potential audience can therefore be much larger and less connected to the original sender.

The key mechanism is automated ranking and distribution. When a system identifies content as relevant or engaging, it may recommend that content to additional users, whose reactions can create further reach. This can turn one upload into a rapidly circulating item rather than a message shared only among acquaintances.

The article does not provide a numerical comparison or a specific audience size. It does state that some content was gaining significant reach through recommendation systems and could cause widespread confusion, apprehension, violence, or disturbance of public order. That potential scale explains MeitY’s focus on amplification.

When misleading material is repeatedly recommended, users may encounter it many times and mistake repetition for reliability. During protests or other public events, this can blur the difference between verified developments and claims presented without context. The article highlights old or selectively edited visuals shown as current, along with misleading claims about emergencies, restrictions, or police action.

The mechanism is amplification. Recommendation systems can give selected content significant reach, while repeated ranking keeps it visible to more users. Audio-visual material depicting law-and-order situations may become especially confusing when it lacks adequate context. Content likely to incite violence can also spread more widely when systems do not identify and escalate it quickly.

MeitY warned that such material could cause widespread public confusion, apprehension, violence, or disturbance of public order. It asked Meta to strengthen rapid identification, review, and escalation, while ensuring that actions remain consistent with Indian law and lawful expression.

The Information Technology Act, 2000, provides the legal framework referenced in the advisory. Under that framework, MeitY highlighted intermediaries’ obligation to exercise due diligence. Social media platforms are therefore expected to take specified care in handling content and operating their services.

The IT Rules, 2021, require intermediaries to inform users about prohibited content. The advisory identifies examples including material likely to cause public disorder or incite a cognisable offence. It also asks Meta to identify and address such content and to review recommendation systems that might amplify misleading or manipulated material.

The ministry further directed Meta to cooperate with law enforcement and other authorised government agencies. That cooperation includes providing information or assistance within the timelines prescribed under the IT Rules, 2021. The advisory also stresses that any action must remain consistent with the Act, the Rules, other applicable laws, lawful expression, and public discourse.

Systems designed to maximize engagement try to predict what will make users click, watch, react, or keep scrolling. Accuracy is not automatically the same as engagement. A dramatic or emotionally intense story may attract faster and stronger reactions than a careful explanation, even when the explanation is more reliable.

The mechanism is ranking. If a system uses interaction signals to decide what deserves visibility, posts producing many comments, shares, or views can receive more recommendations. That extra exposure creates more interactions, which can reinforce the original ranking. Without safeguards, misleading claims can therefore gain momentum through repeated distribution.

The article does not state that Meta’s systems use a particular engagement formula. It does show why MeitY requested reviews of recommendation and ranking systems. The ministry wants misleading, manipulated, or out-of-context material not to be disproportionately amplified, while relevant and credible information receives appropriate diversity and visibility.