Crowd-Sourced Intelligence Agency (CSIA)

The Crowd-Sourced Intelligence Agency (CSIA) is a functional prototype of an open-source intelligence (OSINT) system assembled from publicly available and declassified law-enforcement training materials. It invites the public to take on the role of an intelligence analyst and confront the ambiguity of automated threat classification firsthand.

The program consisted of three main parts: a surveillance interface where users can evaluate Tweets based on their threat to national security, a feature that allowed users to test and search their own social media posts, and a Naïve Bayes supervised machine-learning classifier. Users could review the algorithm’s decisions for accuracy and idiosyncrasies. The CSIA provided first-hand experience with social media monitoring, allowing users to choose how they want to navigate social media surveillance.

The goal of the CSIA was to expose potential problems, assumptions, or oversights inherent in current dataveillance processes to help people understand both the effectiveness of OSINT processing and its impact on our privacy.

Through practice-based research, we found that the public cannot have meaningful oversight over automated OSINT processing systems unless they can develop technical literacies and have access to the data used to train machine learning algorithms. Through the many exhibitions and presentations of this project, we found it highly impactful for members of the public to see how their own tweets are recontextualized in the interface of an OSINT system, and how this can affect how their posts may be interpreted.