The Feed Is Not Neutral: Algorithmic Suppression and the Communities It Was Built to Silence

Before there were platforms, there were griots.

Across West and Central Africa, griots carried history in their bodies—genealogies, treaties, praise songs, warnings, and the names of the dead and the living—and moved it across distances no infrastructure could explain. Information traveled through trust, through relationship, through the ethics of who deserved to hold a story and how it should be told. The community was the network.

We have never stopped needing that. What has changed is who owns the wire.

Social media arrived with the language of democratization. Anyone could speak. Everyone had a megaphone. And for a moment, that felt true. Movements organized in hours. Mutual aid networks formed overnight. Communities separated by diaspora found each other across time zones and borders. The possibilities were real.

So was the backlash—and it was built into the architecture.

What the Data Shows

In 2021, Twitter researchers published a study in the Proceedings of the National Academy of Sciences examining how the platform’s recommendation algorithm shaped political content across seven countries. The findings were plain: the algorithm consistently amplified right-leaning political parties over their left-leaning counterparts — in six of the seven countries studied. It is a finding that raises deeper questions about political identity itself — including whether being a Democrat or a Republican really matters in shaping political outcomes, or whether the partisan frame obscures as much as it reveals.

That same period, internal Facebook research — later disclosed through investigative reporting — found that Instagram accounts that appeared to be Black were approximately 50% more likely to face automatic account suspension than comparable white accounts.

And the documents disclosed by Facebook whistleblower Frances Haugen confirmed what many had long suspected: the platform’s engagement-maximizing algorithm rewarded outrage. The more a post provoked a reaction, the more the algorithm distributed it—and the communities disproportionately targeted by the resulting flood of hate speech and coordinated harassment were communities of color.

This is not a glitch. A glitch is accidental. What these documents describe is a design — a system optimized for engagement metrics that, when left unexamined, encodes and amplifies the same hierarchies that exist off-platform.

#EndSARS and the Cost of a Borrowed Platform

In October 2020, young Nigerians took to the streets to demand the dissolution of SARS—the Special Anti-Robbery Squad, a police unit responsible for widespread torture, extrajudicial killings, and extortion targeting young people, particularly those in creative and tech industries. The movement was organized largely through Twitter, Instagram, and WhatsApp. It was one of the most sustained pro-democracy uprisings in Nigeria’s recent history.

Facebook and Instagram faltered. Facebook’s automated systems flagged posts with the #EndSARS hashtag as misinformation, routing them to a COVID-19 fact-check—the algorithm could not distinguish SARS, the police unit, from SARS, the coronavirus. Bail fund information, livestreams of protests, and evidence of state violence were caught in the same net. Instagram blocked the hashtag. The platforms apologized. The movement continued regardless.

Then came the sequel. When Twitter deleted a tweet by President Muhammadu Buhari in May 2021 — a post that invoked the language of civil war violence against the Igbo people — the Nigerian government banned Twitter outright for seven months. This is the other edge of the same blade. Platforms suppress protest when algorithms misfire. Governments suppress platforms when they don’t.

The people caught between them are the ones whose lives are at stake.

This is what it means to organize on infrastructure you do not own. The landlord can always change the locks.

The Paradox of Black Digital Labor

There is a particular irony that runs through every conversation about social media and Black communities: Black culture is the engine of these platforms’ cultural relevance and profitability, and Black political organizing is treated by those same platforms as a liability to be managed.

The language that shapes online communication — the vocabulary, the cadence, the meme formats, the aesthetic sensibilities — flows disproportionately from Black creators. Studies of viral content consistently show Black cultural production driving engagement across demographics. And yet Black creators have documented, repeatedly and measurably, the experience of having posts suppressed without notice, accounts flagged for content that identical accounts from white users posted without consequence, and reach that collapses the moment content touches anything adjacent to racial justice.

Documents obtained by The Intercept in 2020 revealed that TikTok had instructed moderators to suppress posts from users deemed “ugly, poor, or disabled”—hiding their content from wider audiences to keep the platform aesthetically appealing to advertisers. Black creators later reported the same treatment when posting Black Lives Matter content — sudden, unexplained drops in views and engagement with no mechanism for appeal.

The algorithm does not see a community. It sees a risk variable.

“I Am Because We “Are”—On Platforms That Were Not Built for Us

Ubuntu philosophy holds that personhood is constituted through relationship. I am because we are. The self is not prior to the community; the self becomes real through it. Every griot knew this. Every elder who held a family’s history in memory knew this. A relationship is not incidental to the work—it is the work.

Corporate social media is built on a fundamentally different premise. The individual user is the unit. Engagement—likes, shares, and time on app—is the metric. Community, in the Ubuntu sense, is not something these platforms are designed to cultivate. They are designed to harvest the feeling of community and convert it into data.

This does not mean we abandon these spaces. It means we are clear about what they are and what they are not.

We use them because our people are there. We use them because organizing requires meeting people where they are, and right now, many of our people are on Instagram, on TikTok, on WhatsApp, and on Facebook. But we do not mistake the platform for the village. The village is built in relationship, in accountability, and in the slow and necessary work of showing up—in East Harlem, in Nairobi, in Lagos, in Kampala, and in your own neighborhood—over and over again, whether the algorithm rewards it or not.

What Intentional Engagement Looks Like

If you are using social media as a tool for community and justice, a few things are worth holding:

Know who owns the platform and what they profit from. Elon Musk’s acquisition of Twitter — now X — was not a neutral event for communities that had used that space to document state violence and organize mutual aid. Platform ownership shapes content moderation policy, and policy shapes whose voices survive.

Build off-platform simultaneously. Email lists, text trees, community newsletters, neighborhood gatherings—these are not nostalgia. They are infrastructure that belongs to the people who built it. When a platform changes its algorithm or its terms of service, your email list does not change with it.

Support independent Black and community media. Every click toward an independently owned outlet is an investment in infrastructure accountable to the community rather than to shareholders.

Engage with your body, not just your device. The most durable organizing happens in rooms, not feeds. Social media can announce the meeting. It cannot be the meeting.

And pay attention to what the algorithm suppresses. The content you are not seeing is often more important than what you are. The deeper psychological costs of this environment — the anxiety, the compulsive checking, and the political FOMO it breeds — are examined in our piece on the dark side of online political activity.

The griot did not need a platform to verify the story. The community was the verification. We are still that community. We do not need permission to be heard—but we do need to be clear about which walls we are speaking into and which ones are listening.

---

References

• ‣Proceedings of the National Academy of Sciences — Algorithmic amplification of political content on Twitter across seven countries
• ‣TechXplore — Internal Facebook research on racial bias in Instagram account suspensions
• ‣MIT Technology Review — The Facebook documents disclosed by whistleblower Frances Haugen
• ‣Slate — How Facebook and Instagram mishandled #EndSARS protest content
• ‣Global Voices — The aftermath of #EndSARS, the Twitter ban, and what it meant for young Nigerians
• ‣The Intercept — TikTok moderators instructed to suppress posts from marginalized creators
• ‣Time — Black creators report suppressed reach on TikTok around Black Lives Matter content

Related Reading

• ‣ Epigenetics and Ancestral Memory: What Your Body Remembers
• ‣ Reparations and Healing: A Public Health Argument
• ‣ African Liberation Day: Ubuntu and Ancestral Freedom
• ‣ Mirror Neurons and Communal Healing
• ‣ What Your Feed Won’t Show You: Staying Informed Without Losing Your Peace