Algorithmic Anarchy

The NEET paper leak protests at Delhi’s Jantar Mantar should have remained a legitimate expression of student anger over the collapse of examination integrity. Instead, as the agitation escalated into clashes with the police, abusive sloganeering and an increasingly radical political campaign, Meta’s algorithms emerged as an invisible force multiplying the unrest.

The controversy surrounding the brief removal of Prime Minister Narendra Modi’s message to the youth in the aftermath of the protests only reinforces that concern. Meta has attributed the takedown to a “technical glitch” and apologised. But when the world’s largest social media platform can temporarily suppress the message of the elected leader of the world’s largest democracy during a politically charged moments, the issue extends well beyond a single deleted post.

Who decides what India sees? That question has become impossible to ignore during the Cockroach Janta Party’s protests. Across Instagram and Facebook, users have reported being inundated with CJP videos, reels and protest clips despite never following the organisation or engaging with similar political content. Whether this resulted from coordinated paid collaborations, recommendation algorithms or both deserves a thorough investigation. But the larger democratic concern is that public opinion is increasingly being mediated not by citizens, journalists or elected representatives, but by opaque algorithms designed in Silicon Valley and optimised for engagement rather than democratic responsibility.

Meta’s recommendation engine is not a passive notice board. Every piece of political content that reaches millions has first been selected by an algorithm whose workings remain largely hidden from public scrutiny.

This is hardly unique to India. Around the world, Meta has repeatedly been accused of amplifying polarisation, misinformation and political extremism because outrage keeps users engaged. From elections to ethnic conflicts and episodes of civil unrest, the company has faced persistent criticism that its commercial incentives reward divisive content over balanced discourse.

Democracies cannot afford to outsource the architecture of public debate to corporations whose primary obligation is to shareholders rather than constitutional values. While citizens are entitled to challenge governments and demand accountability, there is an important distinction between a movement that expands because people are persuaded and one that appears to be algorithmically amplified into omnipresence.

Equally disturbing has been the normalisation of abusive political language during the protests. When such content is repeatedly amplified through recommendation systems, platforms cease to be neutral intermediaries and become active participants in degrading democratic discourse.

Platforms that influence elections, protests and public opinion must explain how political recommendations are generated and why particular narratives receive extraordinary amplification.

Silicon Valley companies insist they are merely technology platforms. But their algorithms increasingly exercise editorial power. When software determines which protests dominate national conversations and whose voices disappear, technology has already become politics.