The growing scarcity of critical minerals, coupled with high embodied carbon emissions and persistent pollution from material smelting, highlights the urgent need for a sustainable transformation in materials design. This challenge can be approached as a complex multi-objective optimization problem, requiring the simultaneous consideration of performance, economic viability, recyclability, and full life-cycle environmental impacts. However, the conventional methodologies are increasingly strained by the exponential growth of heterogeneous, high-dimensional data, which significantly constrains their optimization performance in complex engineering scenarios. In response, multi-modal artificial intelligence (AI) offers a transformative pathway by enabling accelerated, data-driven materials design through the integration of diverse textual, visual, and temporal information, thereby efficiently identifying compositions and structures that meet functional and sustainability criteria. This review synthesizes advances across six themes: multi-modal AI foundations for learning composition-processing-structure-property-sustainability relationships; AI-driven sustainable alloy discovery; autonomous laboratories with life-cycle feedback; recyclable and reusable material design; AI-optimized alloys for renewable energy and carbon capture; and data integration challenges, culminating in a roadmap that couples interoperable data infrastructures, human-in-the-loop validation, and autonomous experimentation to accelerate equitable, sustainable materials discovery at scale.
Xu et al. (Wed,) studied this question.