59
眾鑠,積非成是,您相信您的意嗎? 假新聞與社群網分析 Vladimir Kropotov, Lion Gu, Fyodor Yarochkin Senior Threat Researchers @ TrendMicro

Fake News - TW Summit public - s.itho.meœ¾鑠,積非成是,您相信您的意嗎? — 假新聞與社群網分析 Vladimir Kropotov, Lion Gu, Fyodor Yarochkin Senior Threat

Embed Size (px)

Citation preview

  • Vladimir Kropotov, Lion Gu, Fyodor Yarochkin

    Senior Threat Researchers @ TrendMicro

  • HISTORICAL INSIGHTS

  • #hashthinking and short attention span

  • http://ecclesiastes911.net/suggestion/https://www.youtube.com/watch?v=fuIXiXqv978

    Before: good vs bad Now: vs

  • TIMELAW:

  • geolocation:

  • What we observe as FakeNews is about applying old theories to modern means of communication

  • https://www.pdf-archive.com/2016/04/02/30408-1/30408.pdf

    Positive/Neutral vs Negative comments for two candidates during the elections 2015 @QuestionMore TV channel

    Fake news is just a top of the Iceberg

  • Services on the Market (White? Black? Grey?)

  • Connect to friends Read news Collect information to make decisions

    Shopping Investments Votes ...

    Role of social networks

  • Follower service in DeepWeb market

  • Services in DeepWeb market

  • Cross-border availability: Search for follower services (

  • YouTube 7000top 5

  • BandWagon EffectAs more people come to believe in something, others also "hop on the bandwagon" regardless of the underlying evidence.

  • What we expect from Normal Accounts vs Bots

    (1-Norm) ~= Anomaly Now it doesnt works, there is no such line, its kind of fuzzy logic

  • Auto registered accounts with ANtiBan?

  • vs

  • Electronic polls and Voting change.org counts?

  • Properly counting votes in online voting is harder than it seems.

    300000 = 150000

  • Erase Bad stories from the Web

  • The role of public groups

  • Case Studies

  • Federal antimonopoly service threaten to block Facebook, due to Adv of Smoking Blends,

    Facebook will be blocked in Russia in 2013

  • Earning on the collapse of shares - step-by-step guide

  • The story From 2013...

  • Twitter Bots

  • Detecting bot communities

    Accounts profile Avatars Followers Following Account lifetime Registration date

  • Accounts behaviours Posting speed Active hours (how many of you work 24/7) Number of posts per day (no user posts 400 tweets a day at average) Simultaneous activity on promoting topics Timing evaluation (real staff starts slowly) Triggering points (how bots following botmaster) Spotting leading bots (masters) Isolated communities

    Detecting bot communities

  • Convert data to graph (GML) Identify isolated communities Validate retweet/quote only-

    accounts as bots

    Exploring Twitter-Space with Graphs

  • #MacronLeaks caseLast minute swing effect? -

  • #MacronLeaks case

  • #MacronLeaks case

  • #MacronLeaks case

  • #MacronLeaks:

    Strategic leak of information 24 hours before election Domestic media is not allowed to comment Social media is uncontrolled and can be bot-manipulated

  • Donald Trump vs Covfefe

  • Trump Bot Followers

  • Bots

  • Smart bots: cat and mouse game

  • Bot-assisted promotions

  • Bot assisted promotions

  • bots .. bots everywhere :)

  • https://twitter.com/rich_harris/status/957310634378629121?lang=en

    https://www.nytimes.com/interactive/2018/01/27/technology/social-media-bots.html

    Media investigations

  • Twitter handles over the time

  • Compare Following vs Followers

  • TrumpIsVeryBad

  • Political amplification in Bangladesh

  • Social Media Platform business model: - More users -> More Money (?) - More Information -> More Users -> More Money (?) - More Interractions -> More Information -> More Users -> More Money (?)

    Bot activities and services: - create more users - inject more information - create user interractions

    PROFIT?!

    SN and conflict of Interest

  • Related research within Trend Micro

    https://www.virusbulletin.com/conference/vb2014/abstracts/depth-analysis-abuse-twitter/ https://www.trendmicro.de/cloud-content/us/pdfs/security-intelligence/white-papers/wp-an-in-depth-analysis-of-abuse-on-twitter.pdf https://documents.trendmicro.com/assets/white_papers/wp-fake-news-machine-how-propagandists-abuse-the-internet.pdf

    https://www.virusbulletin.com/conference/vb2014/abstracts/depth-analysis-abuse-twitter/https://www.virusbulletin.com/conference/vb2014/abstracts/depth-analysis-abuse-twitter/https://www.virusbulletin.com/conference/vb2014/abstracts/depth-analysis-abuse-twitter/https://www.trendmicro.de/cloud-content/us/pdfs/security-intelligence/white-papers/wp-an-in-depth-analysis-of-abuse-on-twitter.pdfhttps://www.trendmicro.de/cloud-content/us/pdfs/security-intelligence/white-papers/wp-an-in-depth-analysis-of-abuse-on-twitter.pdfhttps://www.trendmicro.de/cloud-content/us/pdfs/security-intelligence/white-papers/wp-an-in-depth-analysis-of-abuse-on-twitter.pdfhttps://www.trendmicro.de/cloud-content/us/pdfs/security-intelligence/white-papers/wp-an-in-depth-analysis-of-abuse-on-twitter.pdfhttps://documents.trendmicro.com/assets/white_papers/wp-fake-news-machine-how-propagandists-abuse-the-internet.pdfhttps://documents.trendmicro.com/assets/white_papers/wp-fake-news-machine-how-propagandists-abuse-the-internet.pdfhttps://documents.trendmicro.com/assets/white_papers/wp-fake-news-machine-how-propagandists-abuse-the-internet.pdf

  • Background Fact checks

    False Flags Agenda

    Conclusion

  • A word of warning~

  • Q & A