The goal of SAM-Smart is to improve the security of smart home devices and networks and to assist users in assessing and dealing with security risks. A key focus is on the development of AI-based security analysis and protective measures. The project follows a user-centered, iterative-incremental approach, aiming to develop, among other things, a voice-based assistance system. This system will help consumers assess specific risks of devices and make secure configurations. Additionally, a privacy dashboard will be developed to enhance data overview and control.

At the core of the project is the "Security & Privacy Gateway," a technical component enabling the detection of IT risks and security incidents while resolving security-related disruptions. It will actively alert users to risks, provide information on security-related topics, and offer guidance. With the help of this gateway, several concepts will be explored and tested, including the development of an "Ambient Security Toolkit" for risk assessment and the creation of a collaborative testbed for pentesting and black- & white-box analysis of smart home devices.

Another focal point of the project is researching and developing automated procedures for detecting security-relevant software flaws and suppressing unwanted information flows. This includes generating hotfixes when manufacturers do not provide updates. Moreover, privacy-preserving distributed machine learning methods for sensor correction will be investigated. The aim is to improve the quality of cost-effective sensors while ensuring data privacy. These procedures will be tested both under laboratory conditions and in real-world environments.

The SAM-Smart project is a cooperation between open.INC GmbH, nuspace GmbH, automITe engineering GmbH, Langlauf GmbH, the University of Siegen and the Institutes of Medical Informatics and IT Security at the University of Lübeck.

This project is funded by the BMBF under grant number 16KISA074.