DOI: 10.36190/2021.12

Published: 2021-06-01
Likes are not Likes A Crowdworking Platform Analysis
Dennis Tatang, Philip Kreißel, Michael Sehring, Florian Quinkert, Martin Degeling, Thorsten Holz

Many Internet users rely on Online Social Networks (OSNs) in their daily lives to read news, find local restaurant recommendations, or learn about products. But lately OSNs have come under scrutiny for allowing misinformation to spread. Compared to the widely discussed problems of fake news and bots, inauthentic likes by real users have not seen much attention in the research community. They have their origin in real accounts that offer likes in exchange for micro-payments through crowd-working platforms. Our analysis of these fake likes from real users is based on almost 90,000 manipulation campaigns managed by a crowd-working company in Germany containing a diverse set ranging from politics to products and services. An additional study on ten users of these crowd-working platforms shows that they merely earn some pocket money and, although they do not like things they politically oppose, do not generally feel responsible for manipulations of recommendations. Our analysis shows that likes should not be trusted and the detection of fake accounts is not sufficient to prevent fake likes in general.