In the energy sector, a forecast can affect operational continuity, people’s safety and the environment.
That is why an algorithm’s accuracy must go hand in hand with the ability to understand its recommendations, assess their uncertainty and stay in control of decisions.
This was the focus of Enrico Zio, Scientific Director of Datrix, in his talk “AI, Energy Systems and Uncertainty” at the G-AI-A: Governing AI for Innovation and Energy Justice conference, held on 8 October at Università Cattolica del Sacro Cuore.
Building reliable AI for energy systems
From predictive maintenance to energy optimisation, building reliable AI means combining data with knowledge of physical systems, making decisions explainable and verifiable, and building in human oversight and operational limits from the design stage.
