Twitter has been experimenting with multiple techniques to improve the integrity and health of its platform. It recently announced that it is making it easier to identify accounts held by heads of state, and also unveiled Birdwatch for a community-driven approach to curbing the spread of misinformation. Today, the company has announced that it has banned hundreds of state-affiliated propaganda accounts from Iran, Russia, and Armenia.
Back in October 2020, Twitter banned 130 accounts from Iran that were disrupting conversations around the first presidential debate. Now, it has banned an additional 238 accounts for manipulation of platform policies. However, the firm has noted that the problematic accounts did not manage to get much engagement.
With regards to its investigations and banning Russian propaganda accounts, Twitter had the following to say:
- Our first investigation found and removed a network of 69 fake accounts that can be reliably tied to Russian state actors. A number of these accounts amplified narratives that were aligned with the Russian government, while another subset of the network focused on undermining faith in the NATO alliance and its stability.
- As part of our second investigation in this region, we removed 31 accounts from two networks that show signs of being affiliated with the Internet Research Agency (IRA) and Russian government-linked actors. These accounts amplified narratives that had been previously associated with the IRA and other Russian influence efforts targeting the United States and European Union.
Lastly, 35 accounts associated with the Armenian government were also banned. These were responsible for furthering the narratives of the Armenian government and also targeted Azerbaijan. Twitter noted that in several cases, these accounts also pretended to be political figures and news operations from Azerbaijan, likely to portray the country in a negative light.
Twitter has announced that it has added information about all of these accounts to its archive for further analysis. This information operations archive was started in October 2018 and currently has the whereabouts of 85,000 accounts from 20 countries which participated in platform manipulation. The Stanford Internet Observatory (SIO) is one of the entities currently studying this data.