Until late 2016, there was very limited awareness that engagement with online misinformation was growing dramatically, due to a lack of visibility about news content consumption online. More recently, there has been extensive work to classify and then map out the misinformation ecosystem but there is yet to be a good way to evaluate if the ecosystem is getting "healthier"—or getting worse. This understanding is crucial for knowing when misinformation is becoming prevalent in order to ensure that there are sufficient resources to address it. It's also important for informing media researchers, policy makers, and technologists who are experimenting with ways to incentivize the production and consumption of accurate news.
To address this need, this project focuses on developing an analysis framework for measuring changes in the news and misinformation ecosystem over time. There are two main challenges with this sort of large scale analysis: automated (and even semi-automated) classification of content is difficult and external parties have limited access to platform or publisher consumption data. This project can use a coarse classification and aggregated data sources to mitigate these challenges enough to provide actionable insights. This work will be structured such that it can be extended as classification approaches improve and as more data becomes available.
Project lead: Aviv Ovadya