International Center for Ethics in the Sciences and Humanities (IZEW)

DAMA: Transparent data autonomy Meta-Assistant

The DAMA project aims to design a smart home environment in such a way that users have an overview of the shared data at all times (transparency) and can control and monitor the functioning of the various assistants and IoT devices themselves (autonomy). The project is carried out in cooperation with the Fraunhofer Institute for Industrial Engineering IAO in Stuttgart. At the IZEW, the ethical evaluation of the technology development is achieved by an operationalisation of the ethical concepts and the accompaniment of the technical implementation.

Duration

August 2020 – July 2023
 

Funding

Baden-Württemberg Stiftung
  

Project description

Artificial intelligence (AI), used among other things in voice assistants and IoT (Internet of Things) devices, controls more and more applications and functions, especially in smart home environments. Users are increasingly losing track of these functions, sensors, and the collection and processing of information that is oftentimes particularly sensitive, not least because it is collected in the domestic environment of one's own home. In the DAMA project, a meta-assistant is to be developed which, based on artificial intelligence, regulates language assistants such as Amazon Alexa, devices in the smart home (e.g. smart refrigerator), but also itself according to personalizable "context levels". The goal is to strengthen the data autonomy and information transparency of the users of smart home applications independently of the application.

The goals of DAMA are:

  •     To ensure informational self-determination and to increase the data autonomy of the users
  •     To improve data literacy of the user in order to ensure well-informed consent to data processing
  •     To develop a context-sensitive regulation of data collection from AI-supported devices and assistants
  •     To develop an overall concept for regulation by a meta-assistant
  •     To develop and implement such a meta-assistant as a prototype.