Can AI become truly sustainable?
Last week our gAIn PI Prof. Dr. Gitta Kutyniok moderated a talk by Prof. Priyadarshini Panda as part of the AI for Good series on Environmentally Sustainable AI.
Her talk explored how spiking neural networks (SNNs) – inspired by the brain’s way of processing information – could open new paths toward significantly more energy-efficient AI systems. By leveraging temporal processing and sparsity, these models have the potential to reduce energy consumption, latency, and computational overhead compared to conventional AI approaches.
This raises an intriguing question: Do we need to move closer to the principles of the human brain to make AI more energy-efficient?
A particularly interesting aspect was also the close integration of algorithms and hardware, highlighting how neuromorphic ideas and algorithm–hardware co-design can translate into practical advances on modern computing platforms.
The recording of the talk is available here:
https://www.youtube.com/live/KPV4tXeXf-U
Date
05th March, 2026
Location
Online