Early Career Scientists Urge Ethical and Inclusive AI for Weather Forecasting

At a time when Artificial Intelligence (AI) is transforming every facet of science and technology, the World Weather Research Programme (WWRP) is ensuring that the next generation of scientists are helping to guide that transformation responsibly. During the 2025 WWRP Scientific Steering Committee (SSC) annual meeting, held on 4 September, early career professionals (ECPs) took centre stage to discuss the promise—and the perils—of AI in weather and climate research.

In a dedicated ECP panel discussion, four young scientists from Argentina, China, Australia, and the United Kingdom called for the development of AI systems that are technically robust, socially relevant, and globally inclusive, urging the meteorological community to pair algorithmic progress with ethics, transparency, and investment in global observation systems.

“The question is not whether AI will transform meteorology—it already is,” said WWRP Co-Chair Professor Tanja Pejovic. “The question is how we can shape it so that its benefits are shared equitably, its predictions are trusted, and its insights are grounded in reality.”

The Role of AI in Weather Research: A Turning Point

Artificial Intelligence is revolutionizing the way scientists analyze climate data, predict extreme weather, and model atmospheric processes. From machine learning models that improve short-term forecasts to neural networks capable of predicting monsoon behavior or heatwaves, AI offers powerful tools to handle the world’s rapidly expanding climate data.

But as the technology evolves, so do the challenges: data gaps, algorithmic bias, ethical accountability, and uneven access to computing resources all threaten to deepen the divide between regions that benefit from technological advances and those that do not.

WWRP, a core scientific initiative under the World Meteorological Organization (WMO), is tackling these questions head-on. Through its focus on inclusive innovation, the programme aims to ensure that new technologies like AI serve both science and society, particularly in regions most vulnerable to weather extremes.

Data: The Foundation for Reliable AI

Panelist Paola Rodriguez Imazio, a meteorologist from Argentina, opened the discussion by emphasizing the importance of data quality and representativeness.

“AI models are only as strong as the data that feed them,” she said. “In data-scarce regions of the Southern Hemisphere, even the most advanced neural networks can produce fragile or misleading results. Without sustained investment in measurement systems, we are building on shaky foundations.”

Rodriguez Imazio’s point resonated strongly with many attending scientists who have long called for greater investment in ground-based and satellite observation networks, particularly across Africa, South America, and the Pacific—regions where climate variability is high but data density is low.

Her remarks underscored a broader truth: AI cannot replace observation. Rather, it depends on a continuous flow of reliable, high-quality data. As climate patterns shift and extreme events grow more frequent, the need for accurate, near-real-time hydrometeorological observations has never been greater.

Equity: Who Benefits from AI in Meteorology?

From his fieldwork in the Himalayas, Rongkun Liu shared firsthand insights into how AI tools can both empower and exclude. His team uses machine learning algorithms to monitor glacial lake outburst floods (GLOFs)—a growing hazard in the region as glaciers melt at an unprecedented rate.

“AI can help us anticipate hazards that were previously impossible to monitor in real time,” Liu said. “But it also risks widening the gap between countries and institutions that have access to powerful technology and those that do not.”

He cautioned that technological dependence—without adequate investment in local capacity—could make vulnerable communities even more reliant on external systems.

For Liu, equity is as important as innovation. “The question is not just whether AI works,” he said. “It’s who gets to use it, who benefits, and who gets left behind.”

His intervention spurred debate on how international organizations, including WMO and WWRP, can help bridge the digital divide by ensuring open data sharing, affordable computing infrastructure, and training opportunities for developing countries.

Trust: Making AI Explainable and Actionable

Representing Australia and the United Kingdom respectively, Dr. Negin Nazarian and Dr. Lewis Blunn showcased how AI can downscale forecasts to city or even neighborhood levels—transforming urban early warning systems.

Their teams are working on AI-driven models that can simulate microclimates, giving city authorities the ability to issue hyperlocal forecasts during heatwaves or severe storms.

“In cities, lives hinge on forecasts that are accurate at the scale of streets and neighborhoods,” Dr. Nazarian explained. “AI gives us that ability—but only if we understand how the model makes its predictions.”

Both Nazarian and Blunn emphasized that trust and transparency are the pillars of responsible AI. “Decision-makers must know why a model predicts what it does before they can act on it,” Blunn said. “Otherwise, we risk creating black boxes that are technically impressive but practically unusable.”

Their work highlights the growing movement within meteorology to develop explainable AI (XAI) systems—models that can not only forecast the weather but also show the reasoning behind their outputs, allowing scientists, policymakers, and the public to make informed decisions.

The Bigger Picture: Policy, Communication, and Ethics

The ECP discussion expanded into broader issues:

• How should AI results be communicated to policymakers who may not be technical experts?

• Can AI help analyze complex systems, like monsoon dynamics, where nonlinear interactions dominate?

• What governance frameworks are needed to ensure that AI is applied ethically and transparently across national meteorological services?

Participants agreed that AI in weather research must remain a human-centered enterprise—guided by values of scientific integrity, inclusivity, and collaboration.

“AI is a tool, not a replacement for human judgment,” one participant remarked. “Our responsibility is to ensure that it augments, rather than replaces, the expertise and experience of meteorologists.”

WWRP’s Vision for an Inclusive Future

The WWRP’s engagement with early career professionals reflects its broader mission: to build a new generation of scientists capable of integrating technological innovation with social responsibility.

By bringing early career scientists into high-level discussions, the Programme ensures that emerging voices help define the standards and principles guiding AI development in meteorology.

The 2025 SSC meeting reaffirmed WWRP’s commitment to:

• Enhancing international collaboration between research institutions.

• Promoting open access to climate data and models.

• Encouraging ethical AI research that benefits vulnerable and underserved communities.

• Developing interdisciplinary training programmes in data science, meteorology, and social science.

“The future of AI in weather research will be shaped not just by algorithms, but by the people behind them,” Bennett concluded. “By empowering young scientists, we’re ensuring that innovation serves both science and society.”

AI and the Future of Weather Prediction

As AI continues to advance, it is becoming clear that its greatest potential lies in collaboration—between humans and machines, and between nations. WMO’s global community of scientists now sees AI not as an end in itself, but as a bridge connecting data, knowledge, and action.

The insights from WWRP’s ECP panel made one thing clear: the future of AI in meteorology depends on inclusivity, trust, and equity. If developed responsibly, AI could dramatically improve the accuracy of forecasts, reduce disaster risk, and save lives—especially in the world’s most climate-vulnerable regions.

As the ECP session concluded, one phrase captured the spirit of the discussion:

“AI will not shape the future of weather research—we will.”