Original title: Don't You Know, Pump it Up! Investigating Cryptocurrency Manipulation in Telegram-Driven Activity
Filipe Moura, Giordano Paoletti, Carlos H. G Ferreira, Jussara Almeida · September 2026
Scammers often use the chat app Telegram to hype up cryptocurrencies and artificially inflate their prices before selling off their holdings. Most studies only look at known scam groups or a few coins, leaving large-scale online hype poorly understood. To solve this, researchers built a system that tracks communication across thousands of public channels. The system filters out irrelevant chat using an encoder modelA type of AI that converts text into numbers so a computer can analyze patterns in the language., spots sudden surges in coin mentions with adaptive thresholdingA technique that automatically adjusts the cutoff for what counts as unusual, based on recent data trends., and links these bursts to market changes using statistical tests. They discovered forty-seven events matching pump-and-dump manipulation and seventy-three sustained market reactions. These manipulative signals showed extreme timing coordination, hitting the chats mere seconds before prices moved. In addition, psycholinguistic checks showed that manipulation messages look completely identical to natural conversations, meaning text clues alone cannot catch them. In total, the detected manipulation events involved over $200 million in financial volume.
Imagine you are in a busy marketplace where some vendors secretly agree to all shout the same price for a fruit at the exact same moment, making everyone think that fruit is suddenly valuable. You try to spot these coordinated shouts by listening to what they say, but their words sound exactly like normal haggling. In this paper, the researchers instead watch the timing — if everyone shouts at once and the price jumps seconds later, it is likely a coordinated scheme rather than genuine demand.
Investors in cryptocurrencies are at risk from coordinated pump-and-dump schemes that use Telegram to spread misleading information. This study shows that these manipulative messages are hard to spot because they sound just like normal discussions, and the schemes can drive price movements within seconds. Regulators and platforms need new detection approaches that look at timing and activity patterns, not just message content.
The abstract acknowledges that its framework only analyzes public Telegram channels, so private or invite-only groups — where much coordination may occur — are not captured. The psycholinguistic finding that pump-and-dump messages are indistinguishable from organic discussions means that text-based filtering alone is insufficient, and the study does not propose a solution to that. Additionally, the quasi-experimental methods can suggest but not prove causation, and the 47 detected events likely represent a lower bound on actual manipulation.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Don't You Know, Pump it Up! Investigating Cryptocurrency Manipulation in Telegram-Driven Activity — Filipe Moura, Giordano Paoletti, Carlos H. G Ferreira et al.
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.