We have a new report on a network that Facebook found & suspended, attributed to individuals associated with Jordanian military. The network had a small presence on TikTok & leveraged Clubhouse discussions. 🧵⤵️
Substantively, the network praised Jordan’s King Abdullah II and the country’s military.
We found 3 TikTok accounts tied to the network. We think this is the 1st known case of TikTok coordinated inauthentic behavior, but it was anti-climatic. The accounts re-posted content from Jordanian Air Force TikTok fan accounts & had low engagement & few followers.
Facebook network assets shared a video which played a recording from a Clubhouse chat about Prince Hamzah, a royal family member who fell out of favor with the King. The video accused foreigners of using Clubhouse as part of “fourth generation social media warfare.”
The network shared other original videos. Many bolstered the state-supported narrative that Basem Awadallah and Prince Hamzah collaborated to destabilize the country.
After Prince Hamzah was arrested in early April, the government banned people from discussing the investigation. We saw a spike in posting during this media blackout, with posts praising the government.
Many posts seems designed to reduce support for ISIS among young people in Jordan, with videos criticizing ISIS’ interpretation of Islam. Note: The network was suspended due to the reliance on fake accounts, not due to the content of the posts.
The network shared posts to out-of-network Groups...sometimes fairly conspicuously.
And oddly several of the suspended profiles had photos of Nick Offerman from his role as Ron Swanson on Parks & Rec.
Last, while we didn’t have visibility into this aspect of the network, Facebook reports that the network was used to harass dissidents.
📢 New research SIO conducted jointly with @Graphika_NYC finds network of suspected Russian personas once again targeting right-wing audiences on alt-platforms. 🧵 1/4 stanford.io/3VVgkmv
@Graphika_NYC The network amplified narratives common within some right-wing communities, such as voter fraud allegations and support for particular candidates, and expanded into pro-Russian arguments regarding the Ukraine war 2/4
@Graphika_NYC Of particular note are fake Kid Rock fan pages across 3 platforms w/posts shared by authentic, big right-wing influencers. Posts promoted unsubstantiated COVID-19 cures & pro-Trump memes. A few hours after our report went out this morning, the Gettr account disappeared. 3/4
It’s a BIG network (>1,000 accounts) that got decent engagement. Its goals included:
1⃣ Promote the work of the Indian Army in India-occupied Kashmir
2⃣ Target specific people perceived as enemies of the Indian government
3⃣ Criticize Pakistan and China
Twitter isn’t publicly attributing the network to any actor, and using open source evidence, we weren’t able to either, but there are some interesting things to note…
First: an oral history documenting the experiences of elections officials during #election2020. It captures the voices of those that guarded American Democracy through a pandemic, massive mis/disinfo & ongoing threats to them, their staff & families. /2
Second: a report that builds on the oral history by focusing on the ongoing threats against 🇺🇸 American elections & the steps that can be taken to further secure elections in an environment where losers will no longer accept that they lost. /3
📢 New work out today from our Tech team & China research team: @joinClubhouse app recently became popular in 🇨🇳. We looked at its data security practices & found a potential risk to mainland Chinese users.
(1) .@AgoraIO, a Shanghai-based startup, provides the backend platform to Clubhouse. This has been widely suspected.
(2/8)
(2) .@joinClubhouse user IDs (not their username — more like a unique serial number) are transmitted in plaintext over the internet, making them trivial to intercept. Chatroom IDs (again, more like serial number) also transmitted in plaintext.
1/ 📢 Out Today: A new report on the contours of Parler. Our team examines growth dynamics (non) moderation policies, and platform growth in 🇧🇷 and 🇸🇦 .
2/ Our analysis pulled data from three snapshots of Parler’s roughly 29 months online. We looked at metrics from the API to map join dates and linguistic patterns on the site.
3/ Parler has roughly 800 moderators, but their moderation practices were purely reactive and did little to filter out spam or fraud accounts.