Abstract

The rapid adoption of Artificial Intelligence (AI) into Smart Home Systems (SHSs) is changing how we engage in our home into continuously sensing and adaptive environments. These systems promise enhance safety and convenience, but they also introduce new forms of vulnerability especially during crises such as fires and burglaries. In the Anthropocene, where environmental disruptions such as (wild)fires are increasing pervasive, smart home technologies must operate under situations of heighted vulnerabilities. Such moments can expose important security and privacy challenges. This research explores how AI-Enabled SHSs exposes or curtails security and privacy during crises scenarios such as fires (and burglaries). Drawing on a mixed-methods approach including mobile application permission analysis and privacy policy evaluation, the study aims to identify mismatches between system behavior and user expectations of privacy. Framed within human-centered computing and ethical theory, this work aims to argue that current smart home architectures insufficiently account for "privacy under stress." It introduces the concept of crisis-aware privacy design, emphasizing the need for systems that maintain trust, transparency, and user control even under degraded conditions. By situating smart home technologies within the broader environmental and ethical challenges of the Anthropocene, this research contributes to ongoing discussions about responsible AI, surveillance, and the future of human-technology relationships.