Researchers at UC Santa Barbara’s National Center for Ecological Analysis and Synthesis (NCEAS) have published guidance for environmental scientists using generative AI in their work. The research, appearing in PLOS Computational Biology, emerged from the team’s experience building the Wildfire Resilience Index.
Key Development
The guidance began as internal NCEAS conversations addressing recurring questions across dozens of research teams: trust considerations, AI access levels, and chat session limitations. Rather than allowing each project to develop independent practices, the center collaborated to establish shared standards.
The Ten Rules Framework
The published guidelines organize recommendations into three phases: pre-coding preparation (including AI selection), active coding practices, and post-coding verification and documentation.
Equity and Access Concerns
The researchers highlight significant disparities in AI adoption benefits. Rachel King notes that “Male researchers report larger productivity gains than their female counterparts,” and UN data shows GenAI usage near 5% in many low-income countries versus two-thirds in some high-income nations. Paid tool tiers may further disadvantage underfunded institutions globally.
Environmental and Economic Impact
The infrastructure supporting GenAI carries substantial costs: data centers could consume 4-12% of U.S. electricity by 2030 and up to 32 billion gallons of water annually by 2028, alongside disruptions to computer science employment.
Journal: PLOS Computational Biology
DOI: 10.1371/journal.pcbi.1014627
Publication Date: August 17, 2026
Source: EurekAlert