Abstract

This disclosure describes how a space with ambient AI, such as a home, an office, a school or a shop, prepares for and responds to natural hazards, including earthquakes, floods, storms, wildfire smoke and extreme heat. The space's house rules include a hazard plan that sets, for each hazard, what may trigger hazard mode, the safe state of every class of device, what extra sensing is allowed to protect life, what responders may receive, and how normal operation resumes. Hazard mode is entered only on an official alert whose signature the hub verifies, on agreement among independent safety-grade sensors on site, or on activation by a member who is present; a message from an app or an AI agent claiming an emergency cannot unlock doors or widen sensing. In hazard mode the hub sends every admitted device its safe state: mobile robots stop or move clear of exits, valves and heaters go to their declared positions, doors are set so that no one can be locked in, agents suspend their own initiatives, and occupants receive clear instructions without any persuasion. Sensing for life safety, such as counting people per room or detecting falls, may widen only for declared purposes, for a limited time, without identifying anyone beyond what rescue requires, and what it captures is erased afterwards except the minimum record of the event. Responders receive a signed, minimal summary of the space. Hazard mode works without power from the grid or a network connection. After the event, every device is self-tested and re-verified before normal operation and the space's conformance mark return. The same plan covers grid stress in extreme weather, in which devices reduce their energy use in a declared order without touching safety functions.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS