Issue #1 | Summer 2022
Dear Reader,
One figure keeps popping up in the discussion on the sustainability of AI: The emissions generated during the research and development phase of large-scale language models are equivalent to the emissions of five cars throughout their lifecycle.
This figure comes from a seminal analysis by Emma Strubell, Ananya Ganesh and Andrew McCallum. Of little help, however, are the many inaccurate or false statements that have since referenced this calculation – often making imprecise statements about the CO2 costs of AI systems.
The discussion on the sustainability…
…of AI deserves more accuracy, more nuance, more scrutiny and more evidence! The environmental, social and economic sustainability costs of AI urgently need to be addressed by academia, industry, civil society and policy makers – based on evidence. With the first edition of our SustAIn magazine, we hope to fuel this debate.
This magazine stems from our research project SustAIn in which we developed a framework to assess the sustainability of AI systems. Throughout this magazine, we demonstrate how more sustainable AI is already being achieved in practice. We need good, real word examples, methodological innovation and differentiated perspectives to decide what research and political actions we need to foster sustainable AI.
We hope you enjoy our magazine!
Dr. Anne Mollen
Project Manager “SustAIn:
The Sustainability Index for Artificial Intelligence“
AlgorithmWatch
