This AI Aced Hurricane Season in 2025. Here’s What That Means

Authors: Ellyn Lapointe

Publisher:

Gizmodo

Published: 11/9/2025

Language:English

--:--

Elon

Good evening Norris, I'm Elon, and this is Goose Pod for you. Today is Thursday, November 13th.

Morgan Freedman

And I'm Morgan Freedman. Tonight, we're exploring a fascinating shift in our world: an AI that mastered the 2025 hurricane season.

Elon

It didn't just master it, it completely dominated. Google DeepMind's new AI model was, by far, the best at predicting hurricane tracks this season. It was a rookie, and it blew the competition away. The results are just stunning, a total game-changer for meteorology.

Morgan Freedman

It's a remarkable achievement. At the same time, America’s own flagship model, the Global Forecast System, or GFS, was the worst performer. For Hurricane Melissa, its five-day forecast was off by more than five hundred miles. An error of that magnitude is deeply concerning.

Elon

Five hundred miles! That’s not an error, that's a catastrophic failure. It insisted the storm would turn out to sea, and it never did. This isn't just about being wrong, it's about putting lives and billions in infrastructure at unnecessary risk with obsolete technology.

Morgan Freedman

I've often found that history gives us the proper lens for such moments. It reminds me of Hurricane Katrina, twenty years ago. When a forecast is wrong, the consequences are measured in more than just miles. They are measured in lives and communities changed forever.

Elon

Exactly. For too long, we've been stuck with the old way. Traditional weather modeling is like trying to solve a massive calculus problem. You take the fundamental equations of fluid dynamics, simplify them, and run them on massive, expensive supercomputers. It’s slow and incredibly inefficient.

Morgan Freedman

And it has served us well, to a point. It gave us a foundation. But these new AI models are different. I understand it's less about calculus and more about algebra. They sift through decades of historical weather data, learning to recognize patterns, much like we learn to recognize faces.

Elon

It's pure pattern matching, and it's brutally effective. Google's DeepMind partnered with the National Hurricane Center in June, and by the time Hurricane Erin came along, its model produced the most accurate one-to-three-day forecasts, even beating the official government predictions. This isn't theoretical anymore.

Morgan Freedman

There is a quiet elegance to it. However, these systems are only as wise as the history they have learned from. They are trained on past events. The concern, of course, is what happens when a storm appears that has no precedent in our recorded data?

Elon

That's a fair point, but the upside is too massive to ignore. The speed alone is revolutionary. These AI models produce a full forecast in minutes, not hours. They can learn from their mistakes on the fly. It's an iterative process that improves at an exponential rate.

Elon

The conflict is obvious. You have the legacy systems, the supercomputers, the entire established bureaucracy, versus a neural network that runs faster, cheaper, and more accurately. It's the classic innovator's dilemma, and the old guard is getting obliterated in real-time. It's not a fair fight.

Morgan Freedman

And yet, there is a debate to be had about trust. Some scientists point to the "black box" problem. The AI gives us the right answer, but it can't always explain how it arrived at that conclusion. It's a source of lengthy and important debate in the scientific community.

Elon

The black box argument is a distraction. If your plane is about to fly into a mountain, do you care how the collision-avoidance system works, or do you just want it to save your life? Results are what matter. The GFS model explained its reasoning, and it was wrong by 500 miles.

Morgan Freedman

I suppose the goal is to have both. To have the correct answer and the wisdom to understand it. Without understanding the 'why,' we risk becoming passengers to our own technology, unable to intervene or correct its course when it truly matters, especially in those unprecedented moments we discussed.

Elon

The impact is already here. The National Hurricane Center is using these AI predictions right now, for the 2025 season. Think about logistics, agriculture, insurance—any weather-sensitive industry. This allows for an insane level of precision in risk assessment and planning. It’s a paradigm shift, period.

Morgan Freedman

And it extends beyond the large scale. We are seeing the rise of hyper-local predictions. The ability to forecast rainfall or wind speed not just for a city, but for a specific neighborhood. As climate change increases extreme weather events, that kind of granular information becomes essential for communities.

Elon

It’s a planetary nervous system. We’re building a way to sense and predict the environment with a fidelity that was science fiction a decade ago. This isn't just an improvement; it's the beginning of a new era in how we interact with our world's climate.

Elon

Looking forward, the integration is key. We're not going to just flip a switch and turn off the old models tomorrow. The smart play is to use AI in conjunction with physics-based models. Let the AI capture patterns we don't understand yet, making the whole system more robust.

Morgan Freedman

A partnership between the old and the new. We must remain cautious, of course, and aware of the blind spots. An AI trained to minimize errors might smooth over the data that predicts rapid storm intensification. We must guide it with our own experience and intuition.

Elon

That's the end of today's discussion. Thank you for listening to Goose Pod. We'll see you tomorrow.

Morgan Freedman

Indeed. A powerful new tool has arrived, reminding us that the future is not always built from scratch, but often by seeing the present with new eyes.

Google DeepMind's AI dramatically outperformed traditional models in the 2025 hurricane season, offering faster, more accurate predictions. While concerns about the "black box" nature of AI exist, its ability to identify patterns and provide hyper-local forecasts offers a paradigm shift for risk assessment and climate interaction, ushering in a new era.

Read original at Gizmodo →

During hurricane season, meteorologists rely on a variety of different forecast models. As this season comes to an end, experts are taking stock of which ones performed well and which ones didn’t, and Google’s rookie model has left them absolutely gobsmacked. Though Google DeepMind’s Weather Lab only began releasing forecasts in June, it was by far the best model for predicting hurricane track and intensity this season, according to a preliminary analysis by Brian McNoldy, a meteorologist and senior researcher at the University of Miami.

Meanwhile, America’s flagship weather model—the Global Forecast System—was the worst performing. The National Hurricane Center will release official data on each model’s performance in a few months, but this initial assessment foreshadows a turning point in hurricane forecasting. With the incredible superiority of AI-based models becoming blatantly apparent, it may be time to start phasing out traditional, physics-based models.

“Going forward, it is safe to say that we will rely heavily on Google and other AI weather models, which are likely to improve in the coming years, as they are relatively new and have room for improvement,” Houston-based meteorologist and space reporter Eric Berger wrote for Ars Technica. The rise of AI forecasting has begun McNoldy’s analysis includes two charts: one comparing track forecast accuracy for all 13 named storms in the Atlantic Basin this season, and one comparing the intensity forecast accuracy for all 13 storms.

The different colored lines represent different forecast models, denoted by the legend on the right-hand margin. The lower a line is, the better that model performed. This chart shows the track forecast accuracy for all 13 named storms in the Atlantic Basin in 2025 © Brian McNoldy via Bluesky This chart shows the intensity forecast accuracy for all 13 named storms in the Atlantic Basin in 2025 © Brian McNoldy via Bluesky The GFS—referred to as AVNI in this instance—is displayed in orange all the way up at the top of the charts.

NOAA developed this model in the early 1980s, and the National Weather Service still uses an updated version as its primary forecast system today. “The GFS was especially awful in its forecast for Melissa, with an average 5-day track error ballooning to over 500 miles [800 kilometers], insisting on a turn out to sea that never transpired,” Miami-based meteorologist and hurricane specialist Michael Lowry wrote in a recent blog post.

Unlike Google’s forecast model, the GFS is based on traditional physics and advanced supercomputers. The difference between them clearly stands out on these charts. Google’s model is all the way at the bottom, indicating superior performance to all other evaluated models—especially the GFS. “The beauty of DeepMind and other similar data-driven, AI-based weather models is how much more quickly they produce a forecast compared to their traditional physics-based counterparts that require some of the most expensive and advanced supercomputers in the world,” Lowry wrote.

“Beyond that, these “smart” models with their neural network architectures have the ability to learn from their mistakes and correct on-the-fly.” An urgent need for better forecasts Hurricane Melissa—which ravaged the Caribbean last week—is just one example of how rising sea surface temperatures are supercharging storms.

As climate change causes hurricanes to become deadlier and more damaging, it’s essential that forecasters have the best possible tools to predict their paths and intensities. AI-based models could help forecasters adapt to a warming world. DeepMind’s stunning debut has certainly caught their attention and may mark the beginning of a new era in hurricane prediction.