90% of online content is paid propaganda, forum user claims

The web has become a board for paid propaganda

The internet is no longer a place for conversation. It has turned into a showcase where someone, with a salary and a script, decides what gets read and what gets buried. The thesis opening this analysis is blunt: in the busiest discussion spaces, the proportion of messages written by community managers serving parties, governments, and propaganda machines vastly exceeds 80% of the content, according to calculations circulating among veteran readers. This diagnosis does not come from an academic observatory, but from those who have spent years reading these same spaces and detecting the pattern.

The phenomenon is not new, but its intensity is. What was once perceived as background noise is now seen as a coordinated choreography: accounts that appear en masse, repeat the same frame, ignore any data that doesn't fit, and disappear when the argument runs out. The result is an ecosystem where spontaneous conversation drowns.

What is denounced: accounts serving a script

The diagnosis unfolds on several fronts. On one hand, spaces dedicated to the war in Ukraine, where some participants are accused of mechanically reproducing Kremlin propaganda, copy-pasting slogans, multiplying identities to simulate consensus, and simultaneously spreading messages in other languages that contradict what they defend in Spanish.

On the other hand, health-themed forums, repurposed as platforms for attacking foreign leaders. Criticism of lockdown policies serves, according to this reading, as a pretext to divert attention toward those who did not apply the same measures. The pattern repeats: a target is chosen, bombarded with out-of-context statements, and any nuance is silenced.

And in the partisan arena, the accusation is more specific. It is claimed that many community managers working for right-wing formations are actually left-wing sympathizers paid to wear down their natural allies. The most cited case is that of a formation which, according to this analysis, saw its digital machinery dedicating itself to attacking another right-wing party instead of its ideological adversaries.

The reward system: thanks when convenient, silence when not

One of the best-described mechanisms is that of selective reinforcement. When a message serves the script, positive reactions rain down. When it contradicts it, emptiness is activated: no replies, no visibility, no trace. It is a cheap and effective way to modulate what is read without needing to delete anything.

Thanks as currency. Those who give them expect loyalty; those who receive them, according to this thesis, become involuntary loudspeakers. The practical consequence is that the average reader does not know who they are talking to. They could be arguing with a neighbor or with an employee with objectives set in an office.

Some clarify that not everything is conspiracy. Part of the most militant messages come from people who are not paid and act out of conviction, replicating frames they have internalized. The distinction matters: an army of wage earners is not the same as a mass of volunteers behaving as if they were.

The cost: a debate that is no longer useful

The accumulated effect is the loss of utility of the space itself. If the majority of content is oriented toward a propagandistic goal, information exchange becomes impossible. Threads drift from a technical topic to an ideological trench in a matter of messages. Conversation breaks.

The paradox is that these same spaces still hold indirect value: they allow reading where parties are heading before they announce it. Changes in discourse, strategic shifts, and internal purges are detected earlier in digital conversation than in press conferences. This is the only return some readers acknowledge for the noise.

---

With this landscape, the prediction is uncomfortable: as long as digital conversation remains a profitable asset for apparatuses, the proportion of interested content will only grow. The reasonable doubt is whether the average reader will ever distinguish human messages from automated ones. For now, the answer offered by this analysis is pessimistic.