SustAIn – Sustainability Index for Artificial Intelligence

The SustAIn project aims to research the sustainability effects of AI technologies, find transformative strategies and promote public and political discourse on the topic. During the project, a sustainability index for AI systems will be created, which will serve as an evaluation system in the future.

Systems based on so-called artificial intelligence (AI) are attributed a growing importance in economy and society. First studies show, for example, that AI systems are associated with a comparatively high energy requirement. The reason for this is the large amounts of data and processor capacity they require. Therefore, with an increased use of AI, the demand for energy also increases significantly. This, in turn, is contrary to the goals of sustainability from an economic and social, but above all ecological point of view.

In the future, AI-based processes should also be able to break new ground in solving ecological problems. However, new technological developments are by no means an end in themselves, but must always be reflected with regard to socially desired developments and offer room for societal learning processes.

The SustAIn project makes a decisive contribution to advancing innovative approaches for the sustainable development of AI-based processes, systematically processing the ecological, social and economic implications of this technological development and carrying them into a broad social discourse. In order to achieve this, civil society (AlgorithmWatch), transdisciplinary sustainability research (Institute for Sustainability Research, IÖW) and AI research and development (DAI-Labor of the TU Berlin) work together on the goal of putting the development and application of AI-based systems in relation to the social goals of sustainability.

The SustAIn project is primarily concerned with the following primary objectives:

  • Systematization of the sustainability effects of AI and gathering of data on these effects (criteria catalog with indicators and concepts for operationalization)
  • Development of an evaluation instrument in the form of a ranking index (Sustainable AI Index)
  • Examination of transformative potentials of applications with a high degree of maturity in sustainability-relevant sectors
  • Development of concrete policy recommendations (policy brief)

To strengthen the social discourse in terms of sustainability, not only dialogue processes with developers – Sustainable AI Labs – are organized, but also recommendations for sustainable AI development – Guidelines for Sustainable AI Development – are formulated. The results of the sustainability assessment – in the form of a multi-dimensional catalog of criteria, a ranking based on it, thematic case studies, and synthesis of the Sustainable AI Labs – are published in annual Sustainable AI Index Reports and discussed with an interested public. A selection of the evaluated, “ranked” applications from the index and other suitable test cases are presented in an interactive format in order to create broad awareness and to enable a discussion on the sustainability effects of AI-based systems.

 

Partners

 

 

 

Funded by

based on a resolution
of the German Bundestag

 

Photo by Chad Davis on Flickr.

AlgorithmWatch is supported by