Towards a framework for harm prevention in web search

We introduce a framework aimed at the information science (IS), information retrieval (IR) and data science communities as well as behavioral and cognitive scientists and policy makers inside government and at corporations operating Web search platforms. The goal of this framework is to instigate collaborative discussion and research across these communities to address potential dangers searchers and society face in modern Web search. We provide an overview of the harms, such as poor health outcomes, and their possible causes, including searcher and system biases being tuned for maximum profit. Modifications to policy, additional evaluation metrics, a mixture of cognitive decision making tools and improvements to the IR system are the suggested pathways. Examples are provided of how the framework can be put into practice.