Artificial Intelligence and the twin transition: the good, the bad, and the ugly

The paper can help you to better understand how AI has the potential both to radically increase and reduce negative environmental effects. Synthesising recent advances, the paper proposes a typology of effects through which AI shapes environmental outcomes: efficiency and footprint effects, prebound and rebound effects, and unlocking and path-escalating effects. The paper highlights how these effects are not properties of AI itself; rather, they emerge based on how firms choose to manage the interactions between effects. To help contextualise the interplay of effects and strategic choices, the paper introduces three AI adoption configurations - sustainability-amplifying, productivity-stabilising, and harm-amplifying AI adoption - and provides actionable guidance for managers, change agents, and policymakers seeking to steer AI deployment towards sustainable effects.