New AI Act Targets Extreme Weather: Federal Agencies Must Use Machine Learning to Forecast Wildfires and Hurricanes

The “Transformational Artificial Intelligence to Modernize the Economy against Extreme Weather and Wildfires Act”—or TAME Act for short—is a major push to drag federal weather forecasting into the age of machine learning. Essentially, this bill mandates that the government, led by the Under Secretary of Commerce for Oceans and Atmosphere, start using Artificial Intelligence (AI) to make weather and disaster predictions faster and more accurate. Within two years, the Under Secretary must create massive, high-quality, AI-ready datasets of historical weather and Earth system data, which are the foundational building blocks needed to train these sophisticated models (Sec. 3).

The AI Forecast: From Hours to Minutes

Right now, traditional weather models rely on complex physics calculations that take supercomputers hours to run. This bill aims to use AI to speed that process up dramatically. The goal is to develop and test AI weather models that can run simulations quickly to check the trustworthiness of current predictions (Sec. 4). Think about the difference this makes in real life: if you’re a construction manager or a farmer needing to know if a severe storm is coming, getting a highly reliable forecast 12 hours out instead of six can save you thousands in damaged equipment or lost crops. The bill specifically requires that the improved weather information be translated into practical advice to help local leaders make better, faster decisions when severe weather is imminent (Sec. 4).

The Wildfire AI Program: Tracking Flames and Smoke

One of the most concrete and immediate changes involves wildfires. Within one year, the Under Secretary must launch a dedicated AI Fire Environment Modeling Program in collaboration with the Secretaries of Interior, Agriculture, and Homeland Security (Sec. 6). This isn't just about predicting if a fire will happen; the AI must be able to forecast how a fire will spread through both wildlands and built-up areas, and critically, detect and track the movement of smoke. For people living near the wildland-urban interface, this means more accurate, faster warnings, potentially saving lives and property. For the firefighter on the ground, it means better tactical information based on real-time data and AI-driven spread forecasts.

Who Pays and Who Gets the Code?

Since this bill requires massive new infrastructure—developing datasets, building new models, and training a federal workforce—it represents a significant investment of taxpayer dollars. The bill acknowledges this by requiring the Under Secretary to develop best practices to minimize the negative environmental impacts of running these energy-intensive AI programs (Sec. 3 & 6). Another key provision involves data access: the bill mandates that the data and code developed under this Act must be released to the public for free under an open license (Sec. 9). This is a huge win for the scientific community and private startups, as they get access to high-quality, government-curated data and code to build their own tools.

However, there’s a catch in the open-data mandate. The Under Secretary has the power to make “necessary changes or accommodations” to withhold data if it threatens national security, intellectual property rights, or the core mission of NOAA to protect people and property (Sec. 9). While these exceptions are reasonable on paper—you wouldn't want trade secrets or sensitive defense data leaked—it does grant the Under Secretary significant discretion over what gets released and what remains restricted. This is a point to watch, as broad discretion can sometimes lead to valuable, non-sensitive data being unnecessarily held back from the public that funded its creation.