Web Surveillance

Drug markets, both on the “surface web” (indexed by search engines) and “darknet” (only accessible through specially configured browsers), will be used as sources for both illicit drugs and new psychoactive substances (NPS). Surface web online stores sell NPS labeled as “not for human consumption,” passed off as plant fertilizers, incense, bath salts, or under other aliases. Almost all are also available on the darknet, and often the NPS market reflects their changing legal status.

Additionally, the content on drug forums can be useful for estimating demographic and temporal trends in drug use, including early and real-time identification of emerging drugs.

In collaboration with the Machine Perception and Cognitive Robotics (MPCR) Lab, housed in the Center for Complex Systems and Brain Sciences at Florida Atlantic University under the direction of Dr. Elan Barenholtz, the Coordinating Center will develop machine learning models that will be trained to identify generic drug-related activity and other substance-specific models to identify particular classes of drugs. In combination, these models will have the capacity to detect drug-related activity that is anomalous and therefore indicative of an emerging trend that may signify the development of novel substances or patterns of use.

The lab will collect drug-related postings and text from the web from: a) online retailers known to be affiliated with commerce of NPS and illicit drugs, and b) drug forum and social media postings. Models will be trained on these data and deployed to identify anomalous drug-related activity which can serve as a source for further investigation and collaboration with NDEWS partners.

The National Drug Early Warning System (NDEWS) is funded by the National Institute on Drug Abuse (U01DA051126) to the University of Florida (PI: Cottler; Co-Is: Goldberger, Nixon, Striley), New York University (Co-I: Palamar), and Florida Atlantic University (Co-I: Barenholtz).