Shadowbans on X: a feature, not a bug

Online free speech advocates breathed a collective sigh of relief in 2022 when Elon Musk bought Twitter, fired entire departments of censors and vowed to make the platform the world’s ‘digital town square’.

👉 Has he made good on that commitment? Well, X’s algorithm just dropped, and it turns out, not really.

Users can now ask Grok if their accounts are penalized or boosted, with the AI tool (around whose architecture X is now powered) informing those who step out of line what they’re doing wrong.

The algorithm reveals penalties for:

🤔 “Conspiracy Theories” – for example, “Epstein-Israel ties, pro-Israel influence on US politics/media”

🤔 “Misinformation Risk” – factual reporting if it’s done with strong editorial slant

🤔 “Anti-Zionism” – criticism of Israel’s policies “as genocide, apartheid, ethnic cleansing, occupation”

🤔 “Dehumanization and Stereotyping” – including “sarcasm toward pro-Israel accounts”

Here’s what the penalties do:

👉 reduce post visibility, deboost replies

👉 “reduced amplification” for non-followers

👉 partial suppression “search_visibility” for “high-sensitivity keywords”

👉 “strong soft shadowban / visibility suppression” for top offenders, including complete scrubbing from search results and posts not appearing even to followers (known as “Shadowed Contrarians,” and ranked by tier)

👉 Downranking for “consistent political/anti-establishment messaging that doesn’t align with dominant or ‘approved’ narratives on either side.”

Meanwhile, the ‘right’ kinds of thought are boosted:

📹 “Topical relevance”

📹 “Alignment with platform trends,” like support for US policies

The algorithm even offers “helpful” advice to offenders, recommending “lower controversy volume” and more neutral engagement, while admitting that some are just lost at this point.

Okay, at least thanks for the honesty, Herr Grok.

More, how it works, complete technical breakdown - to avoid shadowban:

https://www.tweetarchivist.com/how-twitter-algorithm-works-2025

More about algorithm:  https://techcrunch.com/2026/01/20/x-open-sources-its-algorithm-while-facing-a-transparency-fine-and-grok-controversies/

@geopolitics_prime