The Pentagon exception

Lee Fang reported internal records showing that Twitter applied special exemptions after a 2017 CENTCOM request involving 52 Arabic-language accounts. The request presented the accounts as government-attributed. Some later removed those disclosures and adopted ordinary-looking identities; Twitter officials knew of problematic accounts without promptly shutting all of them down. That is concrete evidence of assistance and an enforcement double standard, rather than a theory inferred only from similar talking points. [1]

The broader ecosystem was already visible. Graphika and Stanford’s August 2022 study documented deceptive pro-Western campaigns across several services, including invented personas and media properties. Twitter and Meta supplied removed-account datasets. Importantly, that study did not definitively assign every covert account to CENTCOM. Its findings should not be stretched into a claim that all pro-American online activity was a single Pentagon operation. [2]

The public-interest concern is understandable: people should be able to recognize a government speaking to them, and a platform claiming to police manipulation should explain preferential treatment. The strongest finding concerns protection, concealment and unequal rules. It does not require assuming that every user influenced by these messages was American or that every message was factually false. [3]

The chronology prevents an easy mistake. Twitter and Meta removed overlapping networks in July and August 2022, and the Unheard Voice study appeared in August. Those events preceded Musk’s completed purchase and Fang’s December release. Giving the Files credit for initiating the entire exposure or all account removals reverses the order of events. [1]

The earlier researchers studied covert campaigns across multiple countries and platforms, with uncertain attribution in important areas. Their report also distinguished a known, overt U.S. government messaging initiative from covert activity. That distinction matters: an openly attributed government account is different from an invented local person. Likewise, a disclosed influence effort can deserve criticism without being a fake-identity operation. [1]

Fang’s contribution was to open the company side of the relationship: what officials requested, how Twitter assisted and what staff knew. Contemporary commentary called for investigation and transparency. This guide does not treat that call as proof that a prosecution, compensation program or complete institutional reckoning followed. The documented outcome here is additional public evidence and accountability pressure; the available sources do not establish a Files-triggered criminal judgment against the account operators. [2][3]