According to the company, their software can be used to make forecasts more cheaply than with traditional methods, and it can even predict natural disasters.
Google has developed a new artificial intelligence model, which they claim can predict the weather faster and cheaper than traditional methods, Live Science reports. The software called SEEDS works similarly to generative AI models, which create diverse content based on text, and to large language models, such as ChatGPT.
Forecasting the weather is a complex process that takes into account many factors, some of which can lead to devastating natural disasters such as hurricanes or heat waves. These extreme phenomena are becoming more and more frequent due to climate change, so predicting their occurrence can even save lives. Among other things, Google strives for this with the creation of SEEDS.
Traditional forecasts are made by meteorological services mapping out a number of possible outcomes, and their final “prediction” is an average of these. With this method, they are mostly able to determine the ordinary weather very accurately, however, most meteorological services are no longer capable of dealing with extreme phenomena. The reason for this is that – according to Google’s estimation – to predict a natural disaster with a probability of 1 percent, about 10,000 possible scenarios should be set up, which the services do not have the capacity for.
Artificial intelligence, on the other hand, may be able to work much faster, so in the future it will even be able to predict such events. In addition, according to Google, the cost of operation is negligible compared to the amount of money we spend on traditional methods today.
Tags: Google created weather forecasting artificial intelligence
