THIS month in Antalya, artificial intelligence will have its own day at the 31st Conference of Parties to the United Nations Framework Convention on Climate Change. The Antalya Pledge on AI asks countries to align the rapid growth of AI with global climate goals by addressing AI’s skyrocketing energy demand. The pledge requires AI companies to disclose their data centres’ energy use and shift to clean energy. The COP31 presidency has described the draft pledge as a “political” commitment on how AI is “designed, procured, powered, deployed, measured and managed” in support of climate action. However, the discussion so far has centred largely on how AI is powered rather than on how it can be leveraged to help vulnerable countries adapt to a warming world.
From extreme weather that left 87.5 million people facing acute food insecurity last year to drought draining Middle Eastern reservoirs and deadly heat in South Asia, climate change is impacting harvests, water availability, health, and livelihoods globally. AI can help communities, governments, farmers, and humanitarian organisations anticipate climate-related risks, protect essential resources, and respond faster to disasters. It can proactively convert climate data into actionable decisions that can save lives and reduce losses from adverse climate events.
AI is already proving valuable for climate adaptation. In India’s Andhra Pradesh, a pilot that sent groundnut farmers AI-based advice on the best time to sow by text message in the local language raised average yields by 30 per cent. In Bihar, communities that received AI-based flood alerts faced 30pc lower medical costs after flooding due to earlier evacuations and timely preventative health measures. In Chile, forest sensors analysed by AI detect wildfires about 12 minutes faster than traditional watchtowers. In each case, AI turned data into timely information that people could act on.
Pakistan has also taken steps in this direction. Since October 2025, the Met Office has used MAZU, a Chinese AI-based weather and early-warning system, to improve its forecasts of extreme weather. However, its nine-kilometre resolution is too coarse to capture highly localised weather such as flash floods, and Pakistan’s limited radar network restricts the local data it needs. Without such data and adequate response capacity, AI tools can produce unreliable forecasts.
AI can convert climate data into actionable decisions.
The Antalya Pledge gives Pakistan an opportunity to ensure that adaptation becomes part of the global conversation on AI and climate. To connect the AI pledge to climate adaptation, Pakistan can propose an adaptation-focused lens, with specific language on open climate data, locally adaptable AI models, and shared computing capacity — the building blocks that vulnerable countries like Pakistan need to use AI effectively. Such proposals will carry more weight if they come from many voices rather than just one.
Working through platforms such as the Climate Vulnerable Forum can allow Pakistan to build a joint position with other climate-vulnerable countries and bring their shared priorities to the climate summit’s AI day. For these commitments to be measured, reported, and funded they must be anchored in the wider climate process. Calling for the integration of AI into the Global Goal on Adaptation will ensure that AI-enabled tools are recognised in the indicators adopted at COP30 to track progress and are eligible for the tripled adaptation finance called for by 2035. At home, Pakistan’s multipronged strategy to combat climate change can include: aligning its AI policy’s sector roadmaps with the priorities of its National Adaptation Plan, strengthening its weather monitoring network, improving climate data-sharing across federal and provincial institutions, and training forecasters, researchers and officials to use these tools.
At present, the global scale of planned investment in AI contrasts sharply with the persistent shortfall in climate adaptation finance. Only 32 countries host AI data centres, and Microsoft, Google, Meta and Amazon together plan to spend about $725 billion on AI infrastructure in 2026. For these countries, the pledge’s focus on clean energy matters. But for developing countries, which received just $26bn of more than the $310bn a year needed by 2035 to adapt to climate change, AI’s value as a tool for adaptation matters just as much.
Pakistan comes to Antalya with a clean energy transition shaped by its own citizens, acknowledgement of the need to accelerate clean energy, and an experience of what climate change costs. A pledge that makes AI’s power supply cleaner is a welcome step. One that also puts AI to work helping vulnerable countries adapt to climate change would make it far more meaningful.
Maha Qasim is a climate & sustainability expert. Tahreem Fatima is a sustainability professional.
Published in Dawn, November 5th, 2026
