Purpose The paper aims to examine the environmental challenges of artificial intelligence (AI) in EU law that regard both illicit uses of the technology, i.e. overuse or misuse of AI and its possible underuses. The aim of the paper is to show how such regulatory efforts of legislators should be understood as a critical component of the Green Deal of the EU institutions, that is, to save our planet from impoverishment, plunder and destruction. Design/methodology/approach To illustrate the different ways in which AI can represent a game-changer for our environmental challenges, attention is drawn to a multidisciplinary approach, which includes the analysis of the initiatives on the European Green Deal; the proposals for a new legal framework on data governance and AI; principles of environmental and constitutional law; the interaction of such principles and provisions of environmental and constitutional law with AI regulations; other sources of EU law and of its Member States. Findings Most recent initiatives on AI, including the AI Act (AIA) of the European Commission, have insisted on a human-centric approach, whereas it seems obvious that the challenges of environmental law, including those triggered by AI, should be addressed in accordance with an ontocentric, rather than anthropocentric stance. The paper provides four recommendations for the legal consequences of this short-sighted view, including the lack of environmental concerns in the AIA. Research limitations/implications The environmental challenges of AI suggest complementing current regulatory efforts of EU lawmakers with a new generation of eco-impact assessments; duties of care and disclosure of non-financial information; clearer parameters for the implementation of the integration principle in EU constitutional law; special policies for the risk of underusing AI for environmental purposes. Further research should examine these policies in connection with the principle of sustainability and the EU plan for a circular economy, as another crucial ingredient of the Green Deal. Practical implications The paper provides a set of concrete measures to properly tackle both illicit uses of AI and the risk of its possible underuse for environmental purposes. Such measures do not only concern the “top down” efforts of legislators but also litigation and the role of courts. Current trends of climate change litigation and the transplant of class actions into several civil law jurisdictions shed new light on the ways in which we should address the environmental challenges of AI, even before a court. Social implications A more robust protection of people’s right to a high level of environmental protection and the improvement of the quality of the environment follows as a result of the analysis on the legal threats and opportunities brought forth by AI. Originality/value The paper explores a set of issues, often overlooked by scholars and institutions, that is nonetheless crucial for any Green Deal, such as the distinction between the human-centric approach of current proposals in the field of technological regulation and the traditional ontocentric stance of environmental law. The analysis considers for the first time the legal issues that follow this distinction in the field of AI regulation and how we should address them.