Reporting & Transparency Standards

As the environmental impact of AI grows, so does the demand for transparent reporting. Investors, regulators, customers, and the public increasingly expect organisations to disclose the energy consumption and carbon emissions associated with their AI operations. But standardised reporting frameworks for AI-specific environmental impact are still developing. Existing corporate sustainability reporting frameworks - GRI, CDP, TCFD, and the emerging ISSB standards - cover energy and emissions broadly but don't provide AI-specific guidance. Research initiatives like the AI Energy Star proposal and efforts within standards bodies like ISO are working toward AI-specific metrics and reporting requirements. Some AI companies and research labs have begun voluntarily disclosing the energy consumption and carbon footprint of training specific models, but this is far from universal and the methodology varies. The EU AI Act includes provisions for documenting the energy consumption of AI systems, and broader EU sustainability reporting requirements (CSRD) will require many organisations to disclose environmental data that implicitly covers AI operations. For organisations that want to get ahead of reporting requirements, start by measuring - even roughly - the energy consumption and carbon footprint of your AI workloads. Instrument your training and inference pipelines, work with your infrastructure providers to understand energy sources, and establish a baseline. You can't manage what you don't measure, and the organisations that start measuring now will be better prepared when mandatory reporting arrives.