"These cameras are just like the Eye of Sauron" (arxiv.org)

A recent study highlights the increasing privacy risks associated with AI-driven smart home devices, particularly for domestic workers (DWs) in the UK. Conducted through interviews with 18 DWs, the research reveals that these devices, while intended to enhance home automation, are also tightening surveillance. It introduces a sociotechnical threat model that emphasizes how AI analytics, residual data logs, and cross-household data flows can complicate privacy boundaries. In employers' homes, DWs experience heightened monitoring, while their own usage of smart devices raises unique challenges related to data retention and administrative roles. This investigation is significant for the AI/ML community as it underscores the need to consider the nuanced social implications of AI technology in everyday environments. By framing DW agencies as institutional adversaries, the study calls for improved privacy protections and agency for workers. The findings prompt a reevaluation of how AI functionalities are designed and implemented, advocating for transparency and responsibility in the deployment of smart home technology to mitigate potential sociotechnical threats.