The CIA has been meddling in Twitter’s internal content moderation for years, according to the latest dispatches from Elon Musk’s “Twitter Files” — which also revealed “mountains of insistent moderation demands” from the Democratic National Committee, but not from the GOP.
Two separate threads in the ongoing Elon Musk-sponsored deep dive into the social media’s internal documents were released Saturday by independent journalist Matt Taibbi, documenting how the platform has frequently bowed to government and political pressure.
On June 29. 2020, Taibbi shows, the FBI’s Elvis Chan — who has played a starring role in past Twitter Files releases — asked company executives to “invite an OGA” to an upcoming conference.
“OGA, or ‘Other Government Organization,’ can be a euphemism for CIA, according to multiple former intelligence officials and contractors,” Taibbi explains.
One week later, Stacia Cardille, a senior Twitter legal executive, made the link explicit.
“I invited the FBI and the CIA virtually will attend too,” Cardille wrote to her colleague — and former FBI chief counsel — James Baker on July 8, 2020. “No need for you to attend.”
Baker, one of dozens of ex-FBI agents and executives in Twitter’s ranks at the time, was fired this month for interfering in Musk’s effort to reveal the company’s past transgressions.
From that point, Taibbi writes that “regular meeting[s] of the multi-agency Foreign Influence Task Force (FITF)” — attended by Twitter and “virtually every major tech firm [including] Facebook, Microsoft, Verizon, Reddit, even Pinterest, and many others” — had “FBI personnel, and – nearly always – one or two attendees marked ‘OGA’.”
“Meeting agendas virtually always included, at or near the beginning, an ‘OGA briefing,’ usually about foreign matters,” Taibbi writes.
Through the FITF, US intelligence tasked Twitter analysts with laborious investigations into domestic Twitter accounts alleged to have nefarious foreign connections, the documents reveal — ramping up as the 2020 presidential election approached, but continuing through 2022.
Twitter content monitors analyzed users’ IP data, phone numbers and even weighed whether user names were “Russian-sounding” to confirm the government’s accusations – but often failed to do so.
Taibbi shows how a succession of intelligence reports in 2022 strove to shape news narratives relating to Ukraine and the Russian invasion.
One such report, which lists accounts allegedly tied to “Ukraine ‘neo-Nazi’ Propaganda,’” pushed Twitter to place sites pointing out Hunter Biden’s lucrative role on the board of Busima, the Ukrainian energy company, under a cloud of official suspicion.
Other reports, including one from August 2022, comprised “long lists of newspapers, tweets or YouTube videos” that US intelligence deemed to be guilty of “anti-Ukraine narratives.”
“Intel about the shady origin of these accounts might be true,” Taibbi writes. “But so might at least some of the information in them – about neo-Nazis, rights abuses in Donbas, even about our own government. Should we block such material?”
Meanwhile, a separate thread from Taibbi documented that “Twitter did have a clear political monoculture” — one favoring Democrats.
Democratic Party operatives, and one staffer in particular, barraged Twitter moderators with complaints about Republican memes and spoofs in the run-up to the 2020 election.
In one case, Twitter refused to remove an obviously comical spoof of a “Todos Con Biden” event – in which then-candidate Joe Biden supposedly played a pro-Trump song for a crowd of Hispanic voters.
Moderators also refused to label as “deceptive” a video mashup of Biden repeatedly coughing at a campaign event.
“Because the video is an unaltered excerpt of the Vice President’s speech, our teams consider it to be out of context, but not deceptive,” Twitter told the complaining DNC staffer, Timothy Durigan.
“These rules need revision,” Durigan — the lead analyst of the DNC’s Counter Disinformation Program, according to his LinkedIn account – fumed.
In a carefully courteous response, Twitter sent Durgan what Taibbi calls a “bizarre moderation flow chart” that “showed they can still apply labels to non-deceptive material.”
“If this kind of mechanized speech control can be used one way today, it can be used in another tomorrow, especially if unseen enforcement officials are pushing on the levers,” Taibbi noted.