You are home alone. The heating suddenly turns down. A light switches on in the hallway. A notification appears on your phone. A voice assistant speaks without being asked.
Is it a technical fault, or is someone watching you?
You do not know.
This uncertainty echoes the logic of the Panopticon: never knowing whether you are being watched, you may begin to police your own behaviour. Here, there is no guard in a tower. Surveillance is dispersed through cameras, sensors, logs, alerts, connected accounts, and automated notifications. The home itself appears to watch, predict, and respond.
Smart home devices are marketed as tools of convenience, comfort, security, and care. In abusive contexts, however, they can enable coercive control by sustaining, extending, and concealing it within ordinary domestic life.
So, are we really creating smart homes, or are we creating smart abusers?
The question becomes more urgent when systems do more than obey a direct instruction. What happens when a smart home system learns household routines, predicts what may happen next, and acts without being directly instructed each time?
Fully autonomous smart homes may not yet be commonplace. However, systems are increasingly designed to learn from data, infer patterns, and act on what they learn. These technological advancements could make coercive control more automated, less visible, and harder to escape.
This development calls for precise language to describe the harm it can facilitate. I propose the term ‘Autonomous Systems-Enabled Abuse’ (ASEA) to refer to abuse enabled by AI-driven and autonomous smart home systems (Bukhtiar, 2026).
AI-driven systems learn from data. They may recognise when someone is usually home, detect movement patterns, infer routines, and collect intimate details about everyday life. Autonomous systems can then act on those inferences rather than merely responding to a direct command.
ASEA focuses on what happens when an abuser can exploit both: what a system learns about a survivor, and what it can then do with that knowledge. The smart home can become an element of coercive control when its systems are used to restrict access,
monitor movement, deliver automated messages, change domestic conditions, or anticipate what a survivor may do next.
In this way, smart home systems can extend the reach of coercive control across the domestic environment, allowing an abuser’s influence to permeate multiple spaces and aspects of everyday life. That control can be embedded in settings, accounts, permissions, automation routines, devices, and data flows. The survivor may be dealing with an abusive person and a home configured against her.
This raises an important question: who does the smart home recognise as its legitimate user?
This question points to what I call ‘privilege-by-design’, a term I propose to describe how smart home systems can privilege one user’s authority over others within a shared household (Bukhtiar, 2026).
Smart home systems are often organised around a primary account-holder or administrator with greater control over access, data, recordings, devices, and settings. In some systems, only one “master” user is permitted, leaving other household members with secondary, restricted, or revocable access (Loundens, 2021). Refuge and Avast similarly found that 27% of women surveyed in the UK said administrator access to connected devices was not shared equally or transparently in their household, while 18% had no control over their home Wi-Fi settings (Refuge, 2021).
Homes are shared spaces, and every person within them should be afforded privacy, safety, consent, and autonomy. Yet when administrator access is concentrated in one user, other household members may be unable to alter settings or safely regain control. In abusive contexts, this can enable an abuser to establish or retain disproportionate control over the home.
This is not simply an individual misusing a neutral tool. It is a structural problem.
Smart home infrastructures do not emerge in a social vacuum. Research suggests that, in some households, control over smart home technologies is concentrated among men, while domestic labour falls disproportionately to women (Zeeuw et al., 2021; Hamarowski & Golanska, 2024). Where these unequal divisions exist, designs that concentrate authority in the person who owns, installs, or administers the system may reinforce existing household power imbalances.
Recognising this is essential. Without it, ASEA may be dismissed as a technical fault, isolated misuse, or a private dispute, rather than understood as part of a wider pattern of coercive control.
The broader problem of smart home abuse is already documented. My research asks how AI-driven and autonomous capabilities may extend it. This is not an argument
that every smart home device is inherently abusive. It is a warning against waiting until harm becomes routine before developing safer design, legal safeguards, and effective remedies.
A proactive response begins by naming the harm. ASEA provides language for recognising how AI-driven and autonomous smart home systems can enable coercive control before that harm becomes normalised.
Smart homes should support safety, autonomy, and dignity. If they do not, the question is not whether technology has progressed, but whose power its design has strengthened.
