Artificial intelligence meets LCA: a critical analysis from technical, ethical, and organizational perspectives

Purpose Artificial intelligence (AI) holds strong potential to automate repetitive and data-intensive life cycle assessment (LCA) tasks, thereby reducing costs, enabling scalability, and improving data access. This study critically analyzes the integration of AI into LCA from technical, ethical, and organizational perspectives. It aims to evaluate how AI can be implemented across LCA workflow steps to enhance efficiency without compromising effectiveness, inclusivity, and core LCA ethical values such as transparency and scientific integrity. Methods We conducted a literature review to identify and characterize the technical approaches of current AI-LCA applications across academic research and commercial tools, complemented by a survey assessing current AI uptake in LCA practice and perceived ethical risks. Ethical risks, like accountability or fairness, were derived from existing AI ethical principles and analyzed for their potential conflict with LCA ethical values. Organizational implications including changes in work practices and regional disparities were explored, with attention to the role of AI in social LCA. Identified applications were critically assessed against four categories: technical maturity, potential to enhance LCA practices, operational feasibility, and ethical risks, and used to qualitatively derive criticality of ethical risks. Results and discussion Most AI-LCA applications focus on predicting impact scores and estimating missing inventory data. Few are employing generative AI, while it is mainly used by practitioners for auxiliary tasks like compiling a literature review and report writing. Despite high automation potential, most publicly documented applications remain at proof-of-concept stage, and face operational barriers for inclusivity (coding skills, restricted access) resulting in low adoption. Most critical ethical risks relate to fairness, transparency, accountability and sustainability, driven by black-box models, poor documentation, and representation bias, threatening all core LCA values. Emerging commercial AI-LCA tools present a high potential to widen AI adoption in LCA, but present higher risks toward sustainability, transparency, safety and security. AI is expected to shift LCA stakeholders’ roles from manual tasks to supervising algorithms and validating outputs, while potentially reinforcing regional disparities due to unequal access to AI tools, data, and expertise. Conclusion AI can significantly enhance LCA efficiency and scalability, but without robust collaborative governance, it may undermine methodological rigor and ethical principles. Responsible integration requires need-based adoption, standardized protocols for evaluating AI models in LCA context, human-in-command validation for value choices, transparent documentation, and proactive attention to fairness and inclusivity. These findings provide a foundation for guidance toward effective, inclusive, and responsible use of AI in LCA.