We decode the social effects of media to put their power to good use.


Surfacing the visual malware text-based tracking misses

Memes are increasingly used as vehicles for disinformation and a core strategy in coordinated propaganda and harassment campaigns. As visual tropes that are reproduced, adapted and rapidly spread through social networks, transmitting in-group jokes and veiled meanings, memes can confound journalists, researchers and a public seeking to understand where they came from, what they mean, and the cultural influence they exert. While many researchers and websites, like knowyourmeme, are dedicated to understanding this aspect of visual culture and its practice within various online communities, new contextual search and discovery tools are required to make this work easier, faster, and more responsive.



Vizlab seeks to provide non-specialist, non-technical journalists, researchers and the public visual search and discovery tools to help contextualize and track memes as they travel and evolve online. The project uses computer vision, image forensics, and metadata analysis to build meme family lineages and identify originating source communities online. The project started in March 2018 with automated daily scrapes of visual and associated text and metadata from a number of forums on 4chan and 8chan, including /pol and /leftypol, and has already collected some 100,000 memes.