Abhi Ghadge, associate professor of management of the supply chain at the University of Cranfield in the United Kingdom, says that there was “a kind of general neglect” in terms of climate resilience, although it is starting to change.
The construction of a detailed understanding of a supply chain can however be incredibly difficult, especially for small businesses. Who provides their suppliers? What key raw material is about to become subject to a shortage? Monitoring these details requires long-term commitment and investments, explains Beatriz Royo, an associate professor in the Mit-Zaragoza program in Spain.
Aware of this, the professional service company Marsh McLennan launched a system called Sentrisk last year which, according to her, can automatically analyze the company's shipping manifests and customs clearance recordings to create an image of its supply chain. Sentrisk is based on models of large languages to potentially read billions of PDF documents, according to the customer in question, and automatically trace where the materials and individual parts come. “He could badly read badly something,” said John Davies, commercial director of Sentrisk – although he points out that the system is based on artificial intelligence only to read documents, and not extrapolate beyond them. There is no chance to hallucinate a network of suppliers that does not exist.
Sentrisk combines this analysis of the supply chain with climate risks data in specific locations. “If you need to invest in building a new manufacturing plant, you can perhaps choose a location less likely to be affected by the water shortage,” explains Davies.
Another challenge is that digital twins require a constant update, explains Dmitry Ivanov, professor of supply and operations management chain in Berlin School of Economics and Law. “It's not like a house that you build and the house exists in this form for 100 years,” he says. “The supply chains change every day.”
And although we have a fairly good idea of the way in which climate change will affect the planet as a whole in the coming years, the exact location, the timing and the extent of specific disasters are difficult to predict. This is where new climate risk and extreme weather prediction tools come into play. The semiconductor and the NVIDIA IA giant has a platform called Earth-2, which, in hope, will take up this challenge, with the help of other organizations, including the oceanic and atmospheric national administration.
The idea is to use AI to provide previous warnings of drought or flood, or predict more precisely how a storm will develop. Certain parts of the world have only relatively high information on current weather conditions; Earth-2 uses the same type of AI that sharpens images of your smartphone camera application to simulate high resolution data. “It is really useful, especially for small regions,” explains Dion Harris, principal director of high performance IT and IA factory solutions at Nvidia.
Companies can feed their own data in Terre-2 to further improve predictions. They could use the platform to model climatic and meteorological impacts in specific geographies, but the overall scope of the project is large. “We build the fundamental elements to create a digital twin of the earth,” explains Harris.
