House, Controlled by AI (What Could Go Wrong?)

Date
Type
Workshop
Language
English
Topic
Agent IA
Length
75 minutes

About this talk

This project demonstrates how agentic AI can enable autonomous decision-making and natural-language interaction in smart home systems. By integrating an ESP32 microcontroller with basic sensors (e.g., temperature, motion, or object detection), we show how an AI agent can perceive environmental data, reason about it, and act on user instructions or contextual conditions. For example, the agent can:

  • Adjust the air conditioner temperature to 30°C based on voice or text commands. - Query fridge inventory data (e.g., count remaining eggs). - Detect home occupancy using motion or door sensors.

Photos (1)

Photo of House, Controlled by AI (What Could Go Wrong?)
In the gallery of PyDay BCN (2025), with the post it was published in.