Artificial intelligence (AI), machine learning (ML), and large-scale data infrastructures are shaping environmental and chemical decision-making, accelerating materials discovery, informing regulatory assessments, and guiding industrial innovation. However, without deliberate safeguards, these systems risk reproducing and amplifying long-standing inequities in environmental exposure, data representation, and participation. Cheminformatics—the application of computational tools integrated with experimental data to model chemical structures, properties, and hazards—offers a concrete case study of how environmental data systems can both advance and undermine environmental justice. When datasets, models, or exposure assumptions omit certain geographies, populations, or use contexts, resulting blind spots can reinforce disparities in chemical risk assessment and protection, particularly for communities historically burdened by pollution and adverse health outcomes. In this Perspective, we argue that data equity—encompassing equitable access, representation, governance, and accountability across the chemical and environmental data lifecycle—is essential to ensure that AI-enabled chemistry supports, rather than hinders, sustainable development and environmental protection. We show how biased datasets, unequal access to modeling tools, and opaque decision-making architectures can entrench inequity in digital environmental systems. We then propose a six-element framework—Access, Transparency, Inclusive Design, Capacity, Shared Benefits, and FAIRness (Findability, Accessibility, Interoperability, and Reusability)—to embed equity into environmental and chemical data infrastructures. This framework positions data equity as a foundation for responsible governance of informatics across regulatory science, safer chemical innovation, and community protection, aligned with the UN Sustainable Development Goals.
Fong et al. (Mon,) studied this question.