The triple planetary crisis represents one of the most urgent global challenges, with the IPCC warning of more frequent and severe weather events, rising sea levels and biodiversity loss arising from a 1.5°C rise in global temperatures. Climate change is predicted to cost the global economy $23 trillion annually by 2050. A rise of sea levels by 0.5 to 1 meter by 2100 could cause the displacement of up to 200 million individuals, while climate change could trigger desertification of 25 per cent of the Earth’s land surface by 2050. Global emissions, which currently stand at a record 37.4 billion tons, need to be cut by 45 per cent by 2030 and 57 per cent by 2035 to meet the 1.5°C target.
However, energy demand continues to grow, as do the effects of the energy divide. Although global energy consumption increased in 2023, 685.2 million people worldwide remain without access to electricity. This persistent energy poverty is driven by a combination of factors, including geopolitical instability stemming from global conflicts and the escalating impacts of climate change.
Concurrently, global CO2 emissions are projected to have risen by 0.8 per cent in 2024, a stark contrast to the annual reduction target outlined in the Paris Agreement. Although renewable energy consumption has seen gradual growth, it constitutes only 30 per cent of global energy use, and investments in the energy transition have decelerated in recent years. The IPCC warns that without immediate and substantial action, global temperatures are likely to exceed 1.5°C above pre-industrial levels by the early 2030s, with profound implications for ecosystems, economies and human well-being.
The United Nations Framework on Climate Change (UNFCCC) AI for Climate Action initiative addressed both climate mitigation and climate adaptation, aiming to promote activities aimed at harnessing AI for climate, as well as developing knowledge products that would support policies and best practices in that domain.
In addition, the integration of AI into network technologies holds significant potential to advance sustainable energy consumption and enhance energy efficiency globally. Smart grids can enable dynamic changes in energy supply according to local demand, provide greater accuracy to monitor and forecast energy needs, and enhance load balancing to reduce electricity peaks and costs. 5G serves as a key enabler for smart grids by significantly enhancing their reliability, providing high-speed data transmission and enabling connectivity for a massive number of IoT devices. In a smart grid, household appliances, buildings and entire city regions are interconnected, allowing energy consumption to be monitored and managed more effectively.
An extension of 5G known as 5G-Advanced enables AI algorithms to dynamically adjust energy production and distribution according to real-time demand and increasing energy-sharing opportunities. For example, Huawei, in collaboration with China Telecom and the State Grid Corporation of China, has successfully established a 5G smart grid experimental network in Qingdao. This initiative has set a global benchmark for 5G-enabled smart grids and network slicing, particularly through its innovative application of network slicing technology in power grid management. Similarly, the United Arab Emirates recently announced a $1.9 billion smart grid initiative, developed in partnership with Microsoft’s Co-Pilot, aimed at improving operational efficiency and fostering sustainable energy practices. These efforts have been actively supported by the United Nations and ITU. Notably, the AI for Good platform played a pivotal role in fostering collaborations between technology companies and energy providers, enabling the deployment of AI-driven solutions to optimize energy efficiency and facilitate the integration of renewable energy sources into existing grids.
AI for Good in focus: Leveraging AI in 5G networks to address efficiency challenges
“The energy consumption of 5G base stations is more than three times that of 4G, in order to support enhanced connectivity.” - Guirong Wang, China Telecom.
AI is playing an increasingly transformative role in 5G and telecommunications networks, enabling advanced capabilities such as predictive optimization and real-time traffic management. Autonomous fault detection and anomaly detection features enhance network maintenance by identifying issues early, leading to more robust performance and reducing the need for human intervention. While these advancements improve performance and reliability, they also introduce significant energy demands, underscoring the importance of AI-driven solutions to enhance the energy efficiency of 5G networks and support broader SDGs.
In his presentation “How can AI enable green operation of cloud network infrastructure” at the 2024 AI for Good Global Summit, Guirong Wang, General Manager of the Science and Technology Innovation Department at China Telecom, shared several AI-enabled tools and initiatives that are setting new industry standards for green infrastructure and network operations. Through the 5G Co-construction and Sharing initiative, China Telecom and China Unicom collaborated to build the world’s largest and fastest 5G network, leading to annual electricity savings exceeding 20 billion kilowatt hours. This demonstrates how co-construction and sharing collaborations can enhance network efficiency, reduce costs and accelerate 5G deployment.
As Summer Chen, Vice President of ZTE Corporation, discusses in the AI for Good article “AI’s Role in Digital Transformation,” AI and 5G technologies enable city management systems to collect vast amounts of data, such as gas, water and infrastructure usage to enhance urban safety and performance. Network slicing allows smart cities to create dedicated virtual networks on shared physical infrastructure, enabling tailored connectivity based on real-time demand. This capability supports the dynamic allocation of bandwidth among emergency services, public transport and IoT devices, helping to reduce energy waste and optimize operational efficiency.
ITU’s Focus Group on Autonomous Networks has played a critical role in identifying critical gaps in the standardization of autonomous networks, while also exploring evolution in future networks, real-time responsive experimentation, dynamic adaptation to future environments, technologies and use cases. Key outcomes include its work on concepts such as zero management networks, which provides a fully autonomous network management solution with human oversight, and network orchestration, which automates complex, multi-step processes across diverse network domains and IT systems.
