In recent days, numerous users across social media platforms have complained that videos, reels and posts related to the Cockroach Janta Party (CJP) protests are appearing repeatedly on their Instagram and Facebook feeds despite having no apparent interest in the subject. Many users claim they neither follow CJP-linked accounts nor regularly engage with political content, yet protest-related videos continue to dominate their recommendations.
The complaints have triggered a larger debate about the role of Meta, the parent company of Facebook, Instagram and Threads in determining what political content users see. The discussion is no longer limited to whether a particular campaign has gone viral naturally. Instead, it raises questions about whether recommendation algorithms, engagement-based rankings and moderation policies collectively amplify certain political narratives while restricting others.
Several users on X voiced concerns over what they described as systematic amplification of CJP-related content.
One user alleged that Instagram was being used to push a particular political narrative among young audiences through an organised campaign.
@PMOIndia @narendramodi @AmitShah @HMOIndia
Over the past few days we have seen that there has been a systematic manipulation of algorithms for pushing the CJP narrative on social media platforms of @Meta particularly on @instagram there appears to be an funded campaign for…— 🇮🇳Mohini🇮🇳 (@mohini23451) July 22, 2026
Another claimed that the platform was predominantly promoting anti-police and anti-government content, creating a perception of widespread unrest. Others argued that Instagram’s recommendation system had effectively become an amplifier for CJP protests and urged authorities to investigate.
Some users also alleged that coordinated influencer networks were involved in distributing similar content across multiple accounts. According to these claims, regional coordinators allegedly collaborated with influencers who were paid to circulate protest videos and messaging. Screenshots shared online showed multiple accounts posting similar clips and narratives within a short period.
Have you noticed Instagram flooded with CJP reels lately?
Even if you don’t follow or interact with any CJP accounts, these reels keep appearing in your feed.
That’s because CJP has a network of state-wise social media coordinators who actively collaborate with Instagram… pic.twitter.com/28TQsuhpNN
— Revolutionary Monk (@RevolutionMonk) July 21, 2026
Whether these allegations are independently verified or not, they have reignited scrutiny over how Meta’s recommendation systems function and why political content often reaches users far beyond its original audience.
Meta no longer limits political recommendations
Understanding why such content spreads requires examining how Meta itself says its platforms work.
In 2021, Meta had announced that it would reduce political and civic content across Facebook and Instagram after concluding that many users did not want politics dominating their feeds. During that period, the company also reduced the importance of comments and shares while ranking political posts.
However, this approach changed significantly.
🚨CJP’s Instagram paid content exposed🚨
Same “Delhi police se maar khane ja rahi hu” script, same poses, same dramatic music, different “DM for paid collab”influencers, one clear coordination.
₹5k transferred, reel dropped ✅
You won’t find even a single real protest video… https://t.co/A74gfi4ekh pic.twitter.com/pC1D6DulCU
— Revolutionary Monk (@RevolutionMonk) July 21, 2026
In January 2025, Meta officially announced that it would gradually begin recommending more political content across Facebook, Instagram and Threads through what it described as a “more personalised approach.”
According to the company, recommendations would be based on two broad categories of signals:
Explicit signals, such as likes, follows and reactions.
Implicit signals, such as how long a user watches a post, whether they pause on it, open comments or visit the creator’s profile.
This distinction is important because a user does not necessarily have to support a political movement for the algorithm to conclude that they are interested in it.
Simply stopping to watch a dramatic protest video, even out of curiosity or shock, can become enough for the recommendation system to serve similar content repeatedly.
Algorithms measure attention, not intent
Recommendation algorithms are designed to maximise engagement rather than understand motivation.
For example, a user may watch a protest video because they are trying to understand what happened during a confrontation with police.
Another person may watch because they strongly disagree with the protesters.
Someone else may replay a clip simply to identify a location or verify its authenticity.
From the algorithm’s perspective, however, these differences do not exist.
The platform only records that the content successfully captured attention.
Watch time increases.
Comments increase.
Shares increase.
Profile visits increase.
Collectively, these signals tell the algorithm that the content is engaging, making it more likely to recommend similar posts to more users.
This is why controversial political content often spreads rapidly regardless of whether audiences support or oppose it.
How Instagram decides what appears in explore and reels
Meta’s Help Centre explains that suggested posts on Instagram are selected using multiple factors.
These include:
Previous activity on the platform.
Accounts a user follows.
Information about individual posts.
Popularity of the account sharing the content.
Overall engagement generated by a post.
Similar interests identified through user behaviour.
Suggested political content does not remain confined to followers.
Instead, it appears in:
Home Feed recommendations
Explore page
Reels
Suggested accounts
Consequently, users often encounter political campaigns without actively searching for them.
Research shows algorithms shape political exposure
Academic research has repeatedly demonstrated that recommendation algorithms influence the type of political information users consume.
Research conducted during the 2020 United States presidential election found that replacing Facebook and Instagram’s algorithmic feeds with chronological feeds substantially reduced user activity and changed the political content people encountered.
Although researchers did not observe immediate shifts in political opinions, they concluded that recommendation algorithms significantly alter information exposure.
Another large-scale study analysing aggregated behaviour of nearly 208 million Facebook users found that ideological segregation increased at every stage, from the content users could potentially see to what the algorithm actually displayed and finally to the material users engaged with.
These studies suggest that algorithms do more than organise content. They actively shape the political information environment.
Recommendation policies are corporate decisions
Meta frequently refers to recommendations as algorithmic outcomes.
However, the algorithm itself is designed by the company.
Meta decides:
which engagement signals matter,
how much weight each receives,
what qualifies as recommendation-eligible political content,
and what content is restricted.
Over the past four years, Meta has repeatedly changed these rules.
In 2022, the company reduced the importance of comments and shares while ranking political posts.
By 2023, it shifted greater emphasis towards survey responses regarding meaningful content.
During 2024, Meta restricted proactive recommendations of political posts from accounts users did not follow.
Less than a year later, the company reversed course and resumed recommending personalised political content.
These repeated policy changes demonstrate that visibility is ultimately determined by corporate decisions rather than inevitable technological outcomes.
Why some protest videos gain greater visibility
Another important aspect concerns moderation.
Videos showing police using force during protests are frequently presented as evidence of civil rights violations or excessive use of force.
Such footage can often be defended as legitimate documentation of public events.
By contrast, videos showing protesters attacking police personnel may trigger moderation systems designed to detect violent content.
If organisations associated with separatism or extremism appear in those videos, moderation becomes even more complex because Meta’s Dangerous Organisations and Individuals policy imposes additional restrictions on content involving designated entities.
While Meta allows reporting, criticism and journalistic discussion relating to such organisations, distinguishing between documentation and glorification often proves difficult for automated systems.
The Taliban Case
One of the best-known examples emerged through Meta’s Oversight Board.
An Urdu-language newspaper report discussing the Taliban was removed by Meta despite being legitimate journalism.
Two human reviewers also classified the post as violating platform rules.
The page administrator received a strike and faced restrictions.
The case was later reviewed by Meta’s independent Oversight Board, which concluded that the content constituted permitted news reporting.
Meta subsequently restored the post.
The case exposed two major weaknesses.
First, automated moderation systems failed to understand journalistic context.
Second, the shortage of Urdu-speaking reviewers delayed correction of an incorrect enforcement decision.
The episode demonstrated that moderation errors are not merely theoretical but can directly affect legitimate reporting.
Meta has acknowledged excessive moderation
Meta itself has admitted that its moderation systems have often gone too far.
In January 2025, the company acknowledged that years of expanding moderation systems under political and societal pressure resulted in excessive censorship.
According to Meta, millions of posts were removed every day during late 2024.
The company estimated that between 10 and 20 per cent of enforcement actions may have been mistakes.
Given the enormous volume of daily moderation decisions, even a relatively small error rate translates into millions of potentially legitimate posts being restricted, removed or demoted.
Human reviewers also influence political visibility
Contrary to popular belief, content moderation is not performed solely by artificial intelligence.
Meta employs thousands of human reviewers worldwide.
These reviewers assess: political context, language, satire, journalistic reporting, advocacy, hate speech, extremist references, and potential policy violations.
Human judgement inevitably introduces subjectivity.
Reviewers may interpret political slogans differently.
Language barriers can lead to misunderstandings.
Organised reporting campaigns may also influence which content enters moderation queues.
Consequently, both automated systems and human reviewers collectively shape the visibility of political narratives.
Past allegations of selective enforcement
Questions surrounding Meta’s moderation are not new.
Several publishers have previously alleged inconsistent enforcement of platform rules.
Among them is OpIndia, which has publicly documented multiple disputes with Facebook and Instagram over reduced reach, content removals and restrictions on sharing reports.
The publication has claimed that identical material remained available elsewhere while its own posts were restricted.
It has also alleged repeated removals of content relating to Khalistan-linked organisations, Punjab insurgency and terrorism.
While Meta has not publicly accepted these specific allegations, such disputes continue to fuel broader debates regarding transparency and consistency in content moderation.
Concerns around Meta’s oversight structure
Meta’s independent Oversight Board has also faced criticism from some quarters.
Many have questioned whether the composition of the Board reflects sufficient ideological diversity, particularly regarding issues involving nationalism, separatism, religion and political conflicts.
Reports have previously highlighted links between several board members and international civil society organisations, including institutions that have received funding from George Soros’ Open Society Foundations.
Others have criticised individual members for alleged political affiliations.
Zuckerberg’s admission on government pressure
Another development intensified scrutiny of Meta’s moderation policies.
In August 2024, Meta CEO Mark Zuckerberg informed the United States House Judiciary Committee that officials within the Biden administration had repeatedly pressured the company to remove or suppress certain COVID-19-related content, including satire and humour.
Zuckerberg stated that such pressure was inappropriate and acknowledged that Meta made decisions it would not necessarily make again.
His statement confirmed that governments can attempt to influence platform moderation decisions and that corporate policies may change under external pressure.
The admission became a significant moment in global debates surrounding censorship, political influence and freedom of expression on digital platforms.
The debate extends beyond CJP
Whether the present visibility of CJP protest content is driven primarily by coordinated campaigns, user engagement, recommendation algorithms or moderation policies remains a matter of public debate.
What is beyond dispute, however, is that Meta’s own policies confirm political content is now actively recommended based on user behaviour rather than merely displayed chronologically.
The company’s recommendation systems reward attention.
Its moderation policies determine which posts remain visible.
Its enforcement mechanisms decide what reaches wider audiences and what becomes restricted.
Together, these systems significantly influence how political narratives spread online.
For users who never actively sought CJP protest videos, repeated recommendations can appear as though the platform is promoting a particular political discourse. Meta describes this as personalisation driven by engagement signals. Critics argue that such systems can inadvertently amplify coordinated campaigns and controversial content far beyond their natural audience.
As political communication increasingly shifts onto social media platforms, debates surrounding transparency, recommendation algorithms and moderation policies are likely to remain central to discussions about digital democracy and the influence of Big Tech on public discourse.
