A Promising Direction for Web Tracking Countermeasures

Web tracking continues to pose a vexing policy problem. Surveys have repeatedly demonstrated substantial consumer demand for control mechanisms, and policymakers worldwide have pressed for a Do Not Track system that effectuates user preferences. At present, however, consumers are left in the lurch: existing control mechanisms and countermeasures have spotty effectiveness and are difficult to use. We argue in this position paper that machine learning could enable tracking countermeasures that are effective and easy to use. Moreover, by distancing human expert judgments, machine learning approaches are both easy to maintain and palatable to browser vendors. We briefly explore some of the promise and challenges of machine learning for tracking countermeasures, and we close with preliminary results from a prototype implementation.