The rapid expansion of artificial intelligence has sparked growing debate about whether the technology's benefits can be achieved without creating new environmental and governance challenges. In a recent Nature Communications commentary, an international group of researchers argues that open-source AI could help advance scientific research and support the United Nations Sustainable Development Goals (SDGs), but only if policymakers, developers, and institutions take a more deliberate approach to managing its impacts. The authors, including Min Chen, PhD, and Klaus Hubacek, PhD, professor of environmental economics at the University of Groningen, contend that sustainability considerations must be built into AI development rather than treated as an afterthought.
The team proposes a governance framework intended to align open-source AI tools with sustainability objectives while addressing the environmental, social, and operational challenges associated with large-scale AI systems.
For laboratory leaders increasingly incorporating AI into research workflows, the commentary highlights the importance of evaluating not only the benefits of AI-enabled tools but also the resources required to build, train, maintain, and govern them.
Understanding the environmental footprint of AI
Artificial intelligence is often viewed as a tool for improving efficiency and accelerating discovery. However, the authors note that AI systems depend on extensive computational infrastructure, including data centers, high-performance computing resources, and significant energy consumption.
The commentary argues that organizations should evaluate the sustainability benefits of AI alongside the environmental costs associated with developing and operating these systems. To support this assessment, the authors propose a metric called Return on Environment (RoE).
Unlike traditional measures such as Power Usage Effectiveness, which focus primarily on operational efficiency, RoE is intended to evaluate the broader environmental costs and benefits of AI across its lifecycle. The framework considers factors such as carbon emissions, resource consumption, and the potential societal value generated through AI applications.
According to the authors, broader sustainability metrics could help organizations make more informed decisions about when and how to deploy AI technologies.
Improving resource efficiency through shared infrastructure
The commentary also addresses the growing resource demands associated with training and maintaining increasingly complex AI models. To reduce redundant development efforts, the authors advocate for greater sharing and cooperation across the AI ecosystem.
Lab Quality Management Certificate
The Lab Quality Management certificate is more than training—it’s a professional advantage.
Gain critical skills and IACET-approved CEUs that make a measurable difference.
Rather than repeatedly developing new models from the ground up, organizations could leverage reusable infrastructure and shared model architectures where appropriate. The authors point to techniques such as model pruning and distillation, which previous studies have shown can significantly reduce model size while maintaining performance.
These approaches, they argue, could help lower computational requirements and reduce the environmental burden associated with large-scale AI deployment.
Strengthening governance and accountability
The authors emphasize that technical efficiency alone will not be sufficient to ensure sustainable AI development. Effective governance frameworks, transparency measures, and accountability mechanisms will also be necessary.
The commentary recommends stronger oversight of open-source AI systems through approaches such as third-party audits, standardized reporting, and improved regulatory scrutiny. The authors argue that these measures can help organizations better understand the risks and impacts associated with AI deployment while supporting responsible innovation.
For laboratories and research organizations, these recommendations underscore the growing importance of data infrastructure governance as AI becomes more deeply integrated into scientific workflows.
Interested in lab leadership?
Implications for laboratory operations
Although the commentary is not focused specifically on laboratory environments, its recommendations have implications for organizations that rely on computational research tools. As laboratories increasingly adopt AI for data analysis, modeling, and decision support, leaders may need to consider sustainability metrics alongside traditional performance and cost measures.
The authors conclude that open-source AI can play an important role in advancing sustainable development objectives, but only if its lifecycle impacts are carefully managed. Their proposed framework offers one approach for balancing innovation with environmental responsibility, governance, and long-term resource stewardship.
This article was created with the assistance of Generative AI and has undergone editorial review before publishing.
