Today is Thursday, March 28 2024.
Yesterday we posted about PwC’s 27th Annual Global CEO Survey, summarizing the opinions from over 4,700 CEOs.
Here are two of the several observations from that survey:
Among the megatrends pressuring CEOs to reinvent themselves, none is more important than climate change.
CEOs perceive enormous inefficiencies in a series of routine activities in their companies, representing around 40% of the time spent on these tasks. 60% of CEOs expect that generative artificial intelligence (AI) could help improve this efficiency.
Coincidentally, one day earlier Nature posted an article titled "How AI is improving climate forecasts", with researchers "using various machine-learning strategies to speed up climate modelling, reduce its energy costs and hopefully improve accuracy."
Machine learning is a branch of AI in which computer programs learn by spotting patterns in data sets. This is different than using equations to run simulations and is being more and more considered for weather forecasting and climate modelling. In terms of speed and processing power required - and costs- their results are out much faster - and cheaper - than traditional simulations. On the other hand, machine learnt models still have to prove their accuracy.
In that sense, a few assessment approaches are being undertaken using machine learning:
emulate conventional models
develop foundation models to seek possibly unknown, hidden patterns
hybrid models
The article quotes a few achievements, such as the Australian QuickClim "15 machine-learning models that could emulate 15 physics-based models of the atmosphere", the fast and efficient ACE model developed by the Allen Institute for Artificial Intelligence in Seattle, ClimaX foundation model by Microsoft and University of California, the CliMA project of hybrid modelling, the ‘Digital Twins’ of Earth being developed by NASA and the European Commission, besides an European project called Destination Earth (DestinE).
Two more quotes from the article:
Testing climate models against past climate behaviour is useful, but not a perfect measure of how well they can predict a future that’s likely to be vastly different from what humanity has seen before.
The ultimate goal is to create digital models of Earth’s systems, partly powered by AI, that can simulate all aspects of the weather and climate down to kilometre scales, with great accuracy and at lightning speed.
Click at the image below for this interesting Nature article and references by Carissa Wong.
After all, all these innovative perspectives might well end being very useful for managerial purposes. And practical use, with the support from engineers.
