Thank you to the U.S. Air Force for adhering to the Freedom of Information Act and producing this document about its social media surveillance practices For those curious, the Air Force redacted this citing the FOIA exemption for confidential corporate records/trade secrecy (because it pertains to work done by a private contractor). There's a second page that looks just like the first.

NEW: I've obtained internal Meta docs with examples of posts now allowed under new hate speech rules. Examples include calling children "trannies," “Jews are flat out greedier than Christians," and “immigrants are grubby, filthy pieces of shit” theintercept.com/2025/01/09/fac… The loosened rules allow for overt statements of racial superiority, and unsurprisingly given the national moment, carve out particular latitude for hateful comments about transgender people and immigrants.

NEW: I've reviewed internal Meta records that show the company's dedicated liaison to the Israeli government has repeatedly tried to censor Instagram posts belonging to pro-Palestinian student protest groups and other content related to the war theintercept.com/2024/10/21/ins… Cutler was hired by FB in 2016 as the company's first policy chief for Israel, having previously worked at the Israeli embassy, as a Likud staffer, and advisor to Netanyahu. In a 2020 interview she said "My job is to represent Facebook to Israel, and represent Israel to Facebook"

NEW: U.S. Special Operations Command is seeking the ability to create AI-generated social media users that "Appear to be a unique individual that is recognizable as human but does not exist in the real world" for intelligence-gathering purposes

theintercept.com/2024/10/17/pen… This comes after I discovered last year that SOCOM was similarly interested in creating deepfake propaganda videos theintercept.com/2023/03/06/pen…

Emails I obtained via FOIA request show Elon Musk's Twitter/X was selling user data for government surveillance at the exact same time it was fighting government surveillance in court theintercept.com/2024/03/25/elo… (Of course this contradiction, between professed privacy values and the direct facilitation of surveillance, absolutely applies to Twitter before Musk bought it)

NEW: It turns out Heat Initiative, that brand new organization pressuring Apple to implement its abandoned 2021 proposal to scan everyone's photos for child abuse imagery, is part of a massive US political dark money network theintercept.com/2023/10/01/app… Heat Initiative still won't tell me who's paying its bills, and its website contains zero information about who's involved. But then I noticed something interesting buried in the site's Terms of Use: "THIS WEBSITE IS OWNED AND OPERATED BY Hopewell Fund AND ITS AFFILIATES"

NASA’s Marshall Space Flight Center in Huntsville is the only place in the US I'm aware of that has a statue of a literal Nazi party and SS member proudly on display https://t.co/DVOO9qiww4

https://twitter.com/TimInHonolulu/status/1687976215733563392

impossible to decouple Huntsville's "proud history of aerospace innovation" with the extremely dark history of vast postwar American engineering collaboration with many, many Nazis beyond von Braun

if nothing else AI has turned out to be an incredible powerful tool for surfacing the effects of gutting art education funding and the humanities

https://twitter.com/heykody/status/1662168390352666624

Ever wonder what the Mona Lisa would get into with the cast of Firefly and Pickle Rick?

Google told the public its partnership with Saudi Aramco wouldn't touch Aramco's oil/gas business. I looked into it and that's simply not true. Google Cloud is already being used to help Aramco's methane gas pipelines theintercept.com/2023/02/17/goo… In Sept 2020, Google CEO Sundar Pichai wrote “The science is clear: The world must act now if we’re going to avert the worst consequences of climate change." Two months later, Google announced a deal with Saudi Aramco, one of the greatest contributors to climate change in history

I think one of the biggest recent missteps in tech coverage (and I include myself in this) was uncritically using the term "AI" to mean basically whatever the marketing departments at software companies selling "AI" products want it means. It benefits no one but them. I guess it's not really a "recent" misstep at all, but generally speaking there is nothing "intelligent" about these products, and referring to them as possessing "intelligence" is 1. a huge PR boon for their makers 2. obscures how they actually function