Political propaganda

States and political actors have long used psychological operations to manipulate public perceptions and attitudes. The US military, for example, uses Psychological Operations (PSTOPS) or Military Information Support Operations (MISO) to “create emotions, attitudes, understanding, beliefs, and behaviors favorable to the achievement of United States political and military objectives” (Theohary 2018). Political actors are actively using social media surveillance to segment and target voters based on characteristics such as religion, gender and age. States can use social media surveillance technologies to identify critics of the state, and then to drown out or silence criticism.

Both state and non-state political actors also participate in psychological operations on social media to sway elections and suppress votes. The Berlin-based think tank Tactical Tech reports on the psychological basis of developing psychometric profiling via social media platforms, and how political strategists use such profiling to sway elections (Bashyakarla 2018) (Hankey 2018). The Cambridge Ana- lytica corporation obtained psychometric information on 50 million Facebook users before the 2016 US election (Bashyakarla 2018). In the same time period, Russia created fake Facebook accounts (known as “bots”) that posted divisive and fake news stories with the intent to sway voters towards candidate Trump and away from candidate Clinton. In July 2018, Facebook revealed that it had uncovered Russian attempts to interfere in the November 2018 midterm elections (Fandos and Roose 2018).

In their book on computational propaganda worldwide, Oxford University researchers document that shadowy political forces are using algorithms, automated bots, and human curation to intentionally promote false and misleading information via social media platforms. Computational propaganda is being used to attack political opponents like democracy activists, to sway public opinion, increase polarisation, pollute information ecosystems and manipulate algorithms to falsely inflate “like” counts for political actors or issues in countries across the globe. Computational propaganda was evident during the UK Brcxit referendum, in which politically motivated actors used bots to spread hashtags that seemed to make a small but strategic boost to the “leave” campaign (Wooley and Howard 2018).

Some social media companies assert they have “agnostic” political values. Face- book’s team supports authoritarian leaders in the Philippines, India, Russia and elsewhere who use its platform to threaten journalists, spread disinformation about minorities and undermine the social good (Etter, Silver, and Frier 2017). Facebook offered staff assistance to “embed” with both the Trump and Clinton campaigns during the 2016 US elections. Clinton declined the offer of a Face- book staff support, while Trump’s campaign relied heavily on the daily advice and guidance Facebook staff' gave to their campaign. The Trump campaign also ran three types of ads on Facebook to suppress votes for Clinton. Ads targeted Bernie Sanders supporters, young women and African Americans in key swing states to introduce doubt and convince even a small minority of them not to vote for Clinton (Beckett 2017).

Social media platforms currently do not publish the complete rules for what content they deem inappropriate and what they will remove. Minority groups assert that there appear to be different standards in use for minority groups. For example, Facebook did not remove a Facebook post from a Louisiana Congressional Representative who called for identifying and killing all “radicalized” Muslims in the US but removed a Facebook post and disabled the account of an African American activist who posted a comment noting that all white people are racist (Angwin and Grassegger 2017).