The rapid development of artificial intelligence is leading to increased consumption of electricity and water, creating additional strain on energy infrastructure. This warning was issued by Simon Steell, Executive Secretary of the UNFCCC, during a high-level roundtable on green digital transformation at COP31 in Antalya on November 16.
According to Steell, artificial intelligence has enormous potential for solving climate problems, but its rapid expansion carries significant economic and environmental risks. He emphasized that 'artificial intelligence and other advanced technologies are a double-edged sword. The task is to mitigate their harmful impact, enhance their transformative potential, and use these technologies responsibly.'
The Executive Secretary noted that the growing demand for electricity from data centers creates additional stress on power systems. If this demand is met using fossil fuels, greenhouse gas emissions increase, which hinders the achievement of international climate goals. Furthermore, there is a serious issue of shifting the costs of energy infrastructure onto consumers.
Steell stated that households and businesses should not face higher electricity bills, increased pollution contributing to global warming, and rising inequality in exchange for promises of future technological benefits. He also pointed to instances where data center developers build their own on-site gas power plants to bypass existing grid limitations.
Water consumption is another growing concern. In some countries, local communities are already opposing the construction of data centers due to concerns about water supply, electricity costs, and environmental impact. In light of this, Steell called on technology companies to disclose their energy and water consumption, improve infrastructure efficiency, and power data centers with renewable energy. He stressed: 'Claims of future emission reductions cannot justify increased pollution today.'
At the same time, the head of the UN Climate Secretariat highlighted that artificial intelligence can become an important tool for climate change adaptation. Specifically, AI technologies can improve flood early warning systems, help farmers prepare for droughts, enhance grid efficiency, and facilitate the integration of solar and wind generation.
Steell also stressed the need to ensure developing countries have access to financing, computing power, and reliable data. In his view, the transition to energy is happening at different speeds in developed and developing economies, and artificial intelligence should help bridge this gap, not deepen it. For Uzbekistan, which is expanding its digital infrastructure and renewable energy potential, these issues are of practical importance. The growth of computing power requires careful consideration of electricity availability, grid resilience, and efficient water resource management.
The 'Antalya Declaration on AI' was also presented at COP31—a voluntary framework aimed at reducing the environmental impact of artificial intelligence and promoting AI-based climate solutions.


