Title : Generalized green and red concepts: environmental sustainability of AI, big data, ICT, and general green technologies
Abstract:
Artificial intelligence, big data, and information and communication technologies are becoming central enablers of sustainable development. They support climate modeling, environmental monitoring, smart grids, intelligent transportation, sustainable industry, and digital governance. However, the same technologies are also generating rapidly growing sustainability pressures through electricity consumption, carbon emissions, cooling-water demand, material use, hardware turnover, electronic waste, and broader socio-technical externalities. Therefore, the question is no longer whether intelligence can support sustainability, but how intelligent technologies themselves can be designed, deployed, and governed sustainably.
This talk introduces a generalized framework for understanding green and red technologies in the AI era. Moving beyond the conventional Green AI view, which mainly emphasizes computation, energy, and carbon efficiency, the talk discusses generalized green AI and generalized red AI under a multidimensional Environmental–Economic–Social–Technical/Operational perspective. The discussion is further extended to green and red big data, green and red ICT, and more general green digital technologies. A key argument is that sustainability assessment should distinguish between the service-provision side and the downstream-use side: even if a system appears efficient at the model or infrastructure level, it may still create unsustainable outcomes through excessive data storage, over-provisioned networks, high-frequency hardware replacement, or demand patterns that shift environmental burdens across regions and communities.
The talk will review major sustainability risks in the large-model era, including data-center electricity growth, water consumption, e-waste, and lifecycle impacts. It will then present countermeasures at four levels: metric and accountability reform, ICT design-principle revision and demand shaping, efficiency-oriented large-model development, and energy-infrastructure transition. By connecting conceptual definitions, system-level risks, and actionable interventions, the talk aims to provide a broader view of environmental sustainability and intelligence, and to outline a pathway from narrow green computing toward general green technologies for sustainable development.
