GENEVA: The risk of avalanches can be predicted by artificial intelligence (AI) just as well as by humans, according to the findings of a three-year research project in Switzerland.
The Swiss Institute for Snow and Avalanche Research (SLF) released their report in Davos on Tuesday. The researchers say that machine-trained algorithms and humans have different strengths and weaknesses and can therefore complement each other well.
The SLF has been working with AI models for three years, and found that their predictions are often good. “Sometimes they are clearly not, but even we are sometimes wrong,” said avalanche warning expert Frank Techel.
Algorithms have been used in avalanche prediction for a long time, but what is new is that they now analyse and evaluate the results of many models and provide their own assessment.
Humans have the advantage that, unlike algorithms, they can take into account current observations and feedback from people in the field in addition to data and models when assessing the situation.
On the other hand, due to time constraints, humans can only include the most relevant data in their analyses, whereas computers take all information into account. “The good thing is that the models make different mistakes than we do,” said Techel. – dpa