You may have I recently noticed a drop in AI in your weather app. As businesses rush to infuse artificial intelligence in every productthe wave has come for the modest weather application.
The Weather Company, operator of the Weather Channel, today released a revamped version of its Storm Radar app, featuring an AI-powered weather assistant that lets users customize how they view weather forecasts and maps, switching between layers like radar, temperature, and weather conditions like wind and lightning.
It can also sync with other apps, like your calendar, to send text notifications and weather summaries that combine upcoming weather information with your daily plans. You can stick a voice in there to speak like a former radio meteorologist, if that interests you. Like most weather apps, the data comes from the National Oceanic and Atmospheric Administration (NOAA) and the National Weather Service (NWS).
The app costs $4 per month. It’s only available on iOS at the moment, but the company says an Android version is coming eventually.
“We wanted to create an experience that would get everyone from the casual observer to the seasoned storm chaser up to their level of weather,” says Joe Koval, senior meteorologist at The Weather Company. “If you’re looking for advice on when the weather will be good to walk your dog tomorrow, you no longer need to look at a bunch of disparate pieces of weather data and try to find the answer to that question yourself.”
Of course, you can already check the weather on your phone. Android and iOS devices typically place the weather prominently next to the time. Both Google and Apple have merged their weather apps directly into their smartphones. Since then, AI features have been integrated, offering insights and summaries about the day ahead.
But there are many third-party weather apps, like Storm Radar, Carrot weather, Rain WatcherAnd Weather in Acme-an application of old Dark Sky app creators. New weather apps like Rainbow Weather aim to be AI first. Weather services are also integrated directly into AI chatbots, like Accuweather, which recently launched an application directly in OpenAI’s ChatGPT.
“Everyone has their own idea of what they want in a weather app, what data they are interested in, how they want it to be presented,” explains Adam Grossman, founder of the DarkSky app. “How to create a single weather app that works for everyone? »
DarkSky, one of the most popular iOS weather apps, was purchased by Apple in 2020 and merged with its Weather in Apple service. Grossman eventually left Apple to start Acme Weather, aiming to create a weather forecast service that better reflects forecast uncertainty.
“No matter how good your predictions are, you will be wrong,” Grossman says. “This is something that weather applications have traditionally not done a great job of. Our approach is to try to figure out how to bring those pieces of context back in.”
Weather information repositories typically come from government sources, such as NOAA or other global weather services that collect data from weather satellites, radars, weather balloons, and ground-based instruments. All of this data is integrated into weather forecast models that simulate the physics of the atmosphere. These predictions are often generated by resource-intensive supercomputers, but machine learning models have reduced this processing, making predictions faster. (Although sometimes less accurate, which can be explained by comparing several models.)
Weather applications like Storm Radar and Acme Weather translate this wealth of information by corroborating and compiling models, then helping to create high-resolution maps and visual representation of the data, an area where AI can also be particularly useful.
“Machine learning is probably the biggest change to weather forecasting in some time,” Grossman says. “And they’re just getting started.”
The rise in the use of AI in weather applications comes at a time when government has NOAA dismantled and other federal efforts to track and measure weather conditions, leaving some of the data collection work to private companies. Weather systems also have a harder time predicting extreme weather events and climate disasterswhich are becoming more and more common.
Koval says Storm Radar takes a scientific approach to AI in its application.
“If the NWS issues a warning, the AI will not guess the risk,” says Koval. “It’s going to reference that official warning with your location-specific timeline to let you know how it impacts your plan.”
Storm Radar is a more maximalist approach, with layered complexity similar to something like Google Maps. For the truly weather sick, widgets can be customized to cover the screen, displaying all available weather information. The app’s AI features aim to simplify this data overload, by allowing the AI assistant to give a summary or brief description of the upcoming weather. This can be in text form or through several voices with different accents that aim to sound like TV weathermen.
“You can choose a character ranging from a vintage weatherman to a pop culture fan,” says Koval. “Personalization is really the key to this app.”
Grossman, who says Acme Weather uses AI on the forecasting side, is skeptical of any service (weather or otherwise) that touts its AI just for the sake of having AI.
“It should be transparent; you shouldn’t feel like you’re talking to a chatbot,” says Grossman. “If it’s about surfacing the right content, you need to open it and see what you need to see. You shouldn’t feel like the AI is doing something for you.”
