Social media manipulation: algorithms, bots and computational propaganda
6 likes3,060 views
The document explores the rise of social media as a major news source and the challenges of digital misinformation, including disinformation campaigns, political bots, and the roles of automated accounts in amplifying junk news. It highlights the complex landscape of junk news, public trust issues, and varied industry and governmental responses to combat misinformation. Ultimately, it warns that computational propaganda reflects deeper systemic problems that threaten democracy and suggests that regulatory measures may risk chilling digital discourse.
Social media manipulation: algorithms, bots and computational propaganda
- 1. SOCIAL MEDIA MANIPULATION Center for Media Pluralism and Media Freedom | 26 June 2019 Samantha Bradshaw Oxford Internet Institute Algorithms, Bots and Computational Propaganda Summer School for Journalists and Media Practitioners
- 3. Social media has become a primary source of news around the world
- 4. DIGITAL MEDIA LITERACY THE SPREAD OF DISINFORMATION ON SOCIAL MEDIA DATA-DRIVEN ADVERTISING & CAMPAIGNING FOREIGN INFLUENCE OPERATIONS POLITICAL BOTS AND AUTOMATION SUSTAINABLE JOURNALISM MODELS TROLLS AND ONLINE HARASSMENT DATA PROTECTION AND PRIVACY PLATFORM BLACKBOXES DECLINING TRUST AND INCREASING POLARIZATION WHAT EXACTLY IS THE PROBLEM HERE?
- 5. AUTOMATED ACCOUNTS DATA-RICH PERSONALIZATION ALGORITHMIC AMPLIFICATION PATRIOTIC TROLLING COMPUTATIONAL PROPAGANDA WHAT’S NEW? Gaming algorithms to amplify conspiracy, propaganda, and disinformation about political parties or important political issues. Bot accounts designed to mimic human behaviour to amplify junk content and opinions at the fringe of the political spectrum. Purchasing advertisements and deploying analytics to target citizens with “dark posts” to suppress voter turnout. Paid political commentators who bully, harass and prey on political dissidents or members of the opposition.
- 6. DEFINING DISINFORMATION The study of “fake news” presents several epistemological and methodological challenges: • Encompasses a spectrum of information types; • Limited nuance to binary distinction; • Used to delegitimize political opponents and mainstream media.
- 7. JUNK NEWS TYPOLOGY PROFESSIONALISM STYLEBIAS CREDIBILITY COUNTERFEIT
- 8. JUNK NEWS DOMAINS TOP 29 WEBSITES PRODUCING JUNK NEWS ABOUT US POLITICS, 2016-2019
- 9. JUNK NEWS IN SWING STATES Swing states in the United States shared higher levels of junk news
- 10. 2016 2017 2017 2018 Junk News Shared on Twitter in National Elections, 2016-2018 PROFESSIONAL NEWS: JUNK NEWS
- 11. PROFESSIONAL NEWS: JUNK NEWS EUROPEAN ELECTIONS Low levels of junk overall (< 4%), expect in the Polish sphere (21%) Links to professional news outlets top traffic (34 %) especially in the French & Swedish spheres (>50 %)
- 12. AVERAGE FACEBOOK INTERACTIONS, PER STORY Junk news stories tend to attract more engagement….
- 13. TOTAL FACEBOOK INTERACTIONS …but professional outlets still dominate
- 14. THE PERFORMANCE OF JUNK NEWS DOMAINS ON GOOGLE 0 10,000 20,000 30,000 40,000 50,000 60,000 May'15 Jul'15 Sep'15 Nov'15 Jan'16 Mar'16 May'16 Jul'16 Sep'16 Nov'16 Jan'17 Mar'17 May'17 Jul'17 Sep'17 Nov'17 Jan'18 Mar'18 May'18 Jul'18 Sep'18 Nov'18 Jan'19 Mar'19 0 200,000 400,000 600,000 800,000 1,000,000 1,200,000 1,400,000 1,600,000 1,800,000 2,000,000 May'15 Jul'15 Sep'15 Nov'15 Jan'16 Mar'16 May'16 Jul'16 Sep'16 Nov'16 Jan'17 Mar'17 May'17 Jul'17 Sep'17 Nov'17 Jan'18 Mar'18 May'18 Jul'18 Sep'18 Nov'18 Jan'19 Mar'19 TOP 29 JUNK NEWS DOMAINS MAINSTREAM MEDIA WEBSITES
- 16. WHAT’S UP WHATSAPP? ONE THIRD OF ALL IMAGES SHARED OVER WHATSAPP WERE CONSPIRITORIAL, POLARIZING, OR JUNK ONLY FOUR OF THE 50 TOP SHARED IMAGES ON WHATSAPP WERE FACK CHECKED AS TRUE.
- 18. WHO PRODUCES COMPUTATIONAL PROPAGANDA? HATE GROUPS CONSPIRACY THEORISTS FOREIGN STATES POLITICIANS CLICK BAITERS
- 21. MULTI-DIMENSIONAL CHALLENGES ACROSS MULTIPLE PLATFORMS WITHIN A VARIETY OF COUNTRY CONTEXTS TAKE AWAYS
- 22. Computational Propaganda is a Symptom of much Deeper Problems.
- 23. How Have Social Media Companies Responded?
- 24. Platform Algorithm changes /AI Fact checking and flagging Enhance reporting Human content moderation Enforcement Partnerships and research Media literacy programs Supporting quality journalism Improved transparency Data protection measures Facebook, n= 80 28 19 16 23 42 28 9 8 31 16 Google, n=26 17 9 1 3 6 19 9 15 11 4 Twitter, n=19 8 1 1 4 16 7 3 1 9 1 Total, n= 125 53 29 18 30 64 54 21 24 51 21 CATEGORIZING INDUSTRY RESPONSES
- 25. EMERGENT & REACTIVE TERMS OF SERVICE: ENFORCEMENT ARTIFICIAL INTELLIGENCE CONTENT MODERATION PLATFORM RESPONSES
- 26. GOVERNMENT RESPONSES, 2016-2018 SINCE 2016, MORE THAN 42 GOVERNMENTS HAVE INTRODUCED NEW LEGISLATION DESIGNED TO TACKLE “FAKE NEWS”
- 27. SOCIAL MEDIA PLATFORMS GOVERNMENT RESPONSES, 2016-2018 OFFENDERS GOVERNMENT CAPACITY CITIZENS AND CIVIL SOCIETY Criminalization of “Fake News” Redefinition of Illegal Content Content Take-Down Data Protection Legislation Advertising Transparency Media Literacy and Watchdogs Media Accreditation and Journalistic Controls Parliamentary Inquiries Security and Defence Investments Monitoring and Reporting Portals
- 28. UNINTENDED CONSEQUENCES OF REGULATORY INTERVENTION CHILLING EFFECT on digital public discourse LEARNING CURVE for authoritarian regimes FOCUS ON SYMPTOMS While systemic challenges remain unaddressed
- 29. DEEP THREATS TO DEMOCRACY ARE AHEAD OF US, NOT BEHIND US