Keynote Speaker

“You cannot Predict Electoral Results using Social Media
(but/because Social Media may change the Results)”
By Professor Panagiotis Takis Metaxas, Professor and Chair of Computer Science Department, Wellesley College, USA

Abstract

Ever since the Web and the Social Web appeared, scientists have tried to use tools from statistics and computing to predict the electoral results. Much to their surprise, mathematical tools that are successful in predicting natural events have failed to predict social events, such as elections and the stock market. Researchers have tried to sharpen their mathematical tools but they have failed to realize that prediction only works when the future looks like the past. What is additionally required for the field of computational social science to grow is for engineers to have humility and education in epistemology.

Bio

Panagiotis Takis Metaxas is Professor and Chair of Wellesley’s Computer Science department and a research scholar at the Center for Research on Computation and Society at Harvard University (USA), and the Center for Technology and Global Affairs at Oxford University (UK). His research on the propagation of misinformation online and the un-predictability of electoral results by social media has been extensively covered in the news and honored with several NSF grants and four “best paper” awards.

Takis’ interest in the power of mathematics to discover “the truth” goes back several decades.
His first paper on the topic entitled “Of course it’s true, I saw it on the Internet” was published in 2003, followed by a 2005 paper entitled “Web Spam, Propaganda and Trust”, the first Computer Science paper that was raising the issue of propaganda on the Web and pointing to the dangerous effects of manipulating search engine results.
In the 2010 paper entitled “From Obscurity to Prominence in Minutes: Political Speech and Real-Time Search” they documented the first effort to use Social Media and bots to influence elections in the US. While this work received the Best Paper Award at WebScience 2010 and an NSF grant to develop “TwitterTrails”, a tool for measuring rumor propagation on Twitter, lawmakers did not pay much attention. Unfortunately, the disinformation and propagandistic techniques Takis and his team discovered and warned in “The fake news spreading plague: Was it preventable?” have been used successfully ever since. His current approach to solving the problem is described in a book chapter entitled “Technology, Propaganda, and the Limits of Human Intellect” published this year by MIT Press.