India's Ministry of Electronics and Information Technology has formally asked Meta to re-examine the algorithms that decide what Facebook and Instagram users see, warning that current recommendation systems may be feeding people a narrow, skewed diet of content. The advisory, dated October 10, goes beyond routine content moderation requests. It targets the underlying ranking mechanics that determine visibility, reach and amplification on two of the country's most widely used platforms.
What the Advisory Actually Demands
At its core, the directive asks Meta to ensure its systems do not disproportionately push users toward a single narrative or a narrow set of viewpoints. It also calls for firmer safeguards against content capable of inciting violence or causing public distress. Beyond moderation, the ministry wants structural context built into how media is presented: photographs, videos and graphics tied to public events should carry clear indicators of authenticity, date, location and the circumstances of their creation or sharing. This is a direct response to the way unlabeled or recycled visuals, often stripped of their original context, can spread rapidly and distort public understanding of unfolding events.
The ministry has also instructed Meta to review its content-ranking mechanisms specifically to curb the circulation of manipulated or out-of-context material, and to cooperate promptly with law enforcement and other government bodies when required.
Why Recommendation Systems Matter
Content recommendation engines are not neutral conduits. They are optimization systems, typically tuned to maximize engagement, time spent on a platform, or interaction rates. Because emotionally charged or divisive content tends to generate stronger engagement, these systems can inadvertently reward exactly the kind of material regulators now want curbed. The advisory's concern about users being funneled toward a single narrative reflects a well-documented structural tendency in engagement-driven algorithms, rather than an allegation of deliberate manipulation. The government's intervention specifically during periods of heightened public-order sensitivity suggests particular concern about how such amplification dynamics interact with volatile, fast-moving events.
The Legal Basis and Its Limits
The advisory invokes the Information Technology Act, 2000, and the due diligence obligations under the Information Technology Rules, 2021, which require platforms to inform users about categories of prohibited content, including material likely to incite violence or disturb public order. These rules form the backbone of India's intermediary liability framework: platforms that fail to exercise due diligence risk losing the legal immunity that shields them from liability for user-generated content.
Apar Gupta, founding director of the Internet Freedom Foundation, has challenged the advisory's legal footing, calling it a continuing censorship directive. He has pointed out that terms such as "content without adequate context" and "heightened public-order sensitivity" do not correspond to any defined provision in the Act or the Rules. That ambiguity matters: when regulatory language is broad and undefined, platforms are left to interpret compliance on their own terms, which can lead to overcautious removal of legitimate content or inconsistent enforcement.
What This Means Going Forward
The episode sits at the intersection of platform governance, free expression and public-order management, a tension playing out in regulatory conversations worldwide. Advisories of this kind carry less formal weight than binding rules, yet they still shape platform behavior, since companies operating in India must weigh the cost of non-compliance against the risk of losing safe-harbor protections. Whether Meta makes visible changes to its ranking systems, or simply adjusts enforcement around sensitive events, will be a signal worth watching for anyone tracking how far governments can reach into the architecture of recommendation technology itself.