Personalizing Persuasion Architecture: Privacy Harms and Algorithmic News Media
Date:
For AAAI, I presented research on privacy harms that arise from the personalization of news media.
Abstract
Privacy policies of online news services neglect to disclose what personalization means, the methods used to achieve it, and the risks associated with personalization. This paper analyzes the disconnect between how personal data collection is framed by United States news websites’ privacy policies and how the personalization of persuasive architectures threatens informational and decisional privacy. With set phrases like “personalized recommendations,” “customized user experience,” and “more relevant advertising,” the privacy policies of news websites promote personalization as if it were a wholly beneficent service to the user. This framing is betrayed by the practices of personalization that threaten user autonomy, expose users to undisclosed risks, and subject users to human behavioral research without informed consent. A balance of technological, market, and regulatory solutions is required to mitigate the harms of personalization.