Descripción del proyecto
LIBRA (Leveraging artificial Intelligence to Balance tRade-offs in the digital economy) will assess and compute positive and negative impacts on AI systems on society and the environment and it will optimize them to support policy-makers in Europe and beyond. Artificial Intelligence (AI) contributes to the development of the digital economy and expands the opportunities for sustainable development. However, as computing needs grow, AI systems and applications raise concerns about their environmental implications. Acknowledging this trade-off and achieving an optimal combination between the carbon footprint of AI systems and the benefits they contribute spurring is becoming a key policy concern, but remains an unresolved issue. By combining finance, AI and decision theory, LIBRA aims at filling this gap developing, testing and proposing policy-oriented innovative approaches and tools to: (i) map how investments in AI systems and applications contribute to generating carbon emissions worldwide; (ii) discover how the digital economy enables the achievement of the 17 Sustainable Development Goals (SDGs); (iii) compute the optimal balance between carbon emissions and sustainability objectives to inform national policies. LIBRA will contribute to the ongoing debate about AI and sustainability by combining approaches from computer science (i.e., Natural Language Processing and Deep Reinforcement Learning), complex systems (i.e., network science) and economics (i.e., decision theory) in a holistic framework. The project outcomes will support the discussion around the EU AI Act and the development of a solid sustainability roadmap which accounts for the growing opportunities spurred by the digital economy. LIBRA will last 24 months and it will be mainly conducted at the Royal Institute of Technology (KTH) in Stockholm (Sweden) with a secondment period at the University of Zurich (UZH) in Zurich (Switzerland).