How do households renegotiate voice assistant use over time?

Alizadeh, Lee, Randall, Tolmie, Pins and Stevens report a three-year longitudinal study (January 2020–December 2022) of 30 households (33 participants, 23 with Alexa, 14 with Google Assistant, 3 with both) in Germany, combining iterative interviews, photographs, home maps and a three-week diary study. The authors find that "domestication unfolded not as a linear sequence of stages but as a dynamic process in which devices were either embedded into routines, withdrawn from use, or repurposed in response to changing circumstances." Across trajectories, participants attributed four recurring roles to their voice assistants: "(1) negotiators, (2) separators, (3) mediators, and (4) amplifiers of shared life." The authors name this pattern a "domestication calculus," capturing how spatial, relational and routine dimensions evolve together rather than in isolation.

The paper's mechanism is that life transitions — not novelty decay alone — drive renegotiation of a voice assistant's place in the home. A participant who moved out of a shared flat with her partner stopped placing the device in communal areas, "despite having previously shared it freely with flatmates," once privacy concerns attached to living alone. Spatial trajectories also diverged by generation: older participants "tended to expand the presence of VAs across multiple rooms," while some younger participants "gradually marginalized the device, placing it near the TV, on shelves, or in visually peripheral locations, treating the VA more as a decorative or dormant object." The authors contrast this with shorter studies — such as Cho et al.'s twelve-week, five-stage model (preadoption through acceptance) — which frame declining use after novelty wears off as a terminal endpoint rather than one branch among embedding, withdrawal and repurposing.

This extends the trajectory logic found in Do chatbot relationships lose their appeal as novelty wears off? and How does emotional chatbot use develop from practical use? from conversational chatbots into a physically situated device: where those studies track novelty fading or disruption within the human-AI dyad, Alizadeh and colleagues show the disruption is often exogenous — a move, a breakup, a child growing older — reshaping where and how the device is used. The domestication calculus also parallels Why do users respond so differently to AI chatbots? in treating context and time as structuring variables rather than noise, though the VA study grounds this in physical placement and household routine rather than a chatbot's conversational ecology.

The study does not establish how these dynamics generalize beyond its sample: the authors note the age distribution was uneven, non-primary household members were scarcely represented (three cases only), and several participants already owned a voice assistant before the study began, which may have shaped their starting expectations. All data collection and translation occurred within German households, so the four roles and the calculus itself are offered as "trends discerned from this sample," not population claims. The implication the authors draw is correspondingly modest: designing for long-term domestication means supporting repurposing and renegotiation across life transitions, not just smoothing early adoption.