Some people think, that the evolution of Artificial Intelligence can lead us to a global crisis. In my opinion, it is quite opposite! Artificial Intelligence is a very valuable member of the crisis management team right now and it is ready for new challenges. And nowadays there is no shortage of those …
In crisis management pressure is huge. Theoretically, the recipe for success is simple: taking the right steps in the shortest possible time. Therefore, in practice, success depends on many factors. First of all, professional crisis management is impossible without detecting, data analysis, and forecasting. It sounds like a job for Artificial Intelligence, doesn’t it?
Flood warning system
One of the most famous predictions of natural disasters was made by Babb Vanga. The Balkan Nostradamus, who died in, 1996 predicted the tsunami in Thailand (2004). However, Asia and Pacific is not the only region threatened by natural disasters. Demand for flood forecasting is high also in Europe.
On July the 10th (four days before the disaster) European Flood Awareness System sent out flood warnings to the german administration. Such a rapid response would not have been possible without TSAR AI – an AI-driven emergency management platform. Namely, Artificial Intelligence enables automatic localization of areas at risk of natural disasters. Even the smallest changes in the water bodies will not slip its notice and will be analyzed in the context of flooding risk.
Of course, a well-functioning flood warning system is only half the battle. Finally, during the floods in Germany died at least 188 people. Public opinion blames for it primarily communication problems between the citizens and the state. Perhaps, in the near future, this gap will be filled by the use of AI. Some scientists research the application of machine learning in disaster risk communication. One of the more interesting initiatives in this area is ITU Focus Group on ‘AI for natural disaster management’.
Early warning system bankruptcy
Due to the COVID-19 pandemic, not only a wave of floods but also a wave of bankruptcies could be a problem, which the world will have to face. History of the global financial system proved that even the biggest companies are vulnerable to insolvencies. Enron, Lehman Brothers, Wirecard – these stories shook the whole world. Increasingly, experts point out, that approach of Artificial Intelligence can prevent similar collapses.
An early warning system should be able to reveal enterprise threatens. The most obvious way is to identify them by financial data. Existing models base there on analysis of financial statements. The most important of these models are: Logistic Regression (LR), Multivariant Discriminant Analysis (MDA), Ensemble method, Neural Networks (NN), Support Vector Machines (SVM), Deep Belief Network (DBN), and Convolutional Neural Network (CNN). Part of these bankruptcy prediction models base on machine learning and part on deep learning methods. As you can see, the researches on this topic are extensive but searching for the perfect solution is still ongoing.
Unfortunately, not always reliance on financial data is a good choice. Financial statements are sometimes not entirely reliable, as evidenced by Wirecard scandal. To make accurate predictions, it is necessary to analyze also non-financial data associated with the company and its environment. Eagle Eye seems to be an interesting project in this field.
Eagle Eye is a predictive model, that analyses data available on the internet. How it actually works? Machine learning techniques enable analyze of information related to the company or its environment. Eagle Eye is able to correlate worrying signals and recognize certain patterns. Jan Balatka, Partner of Deloitte believes that: “Only AI can handle the vast volumes of data on the internet and find correlations between signals and credit risk that humans would not even think of. Once we find certain patterns, Eagle Eye constantly monitors the internet to look out for them”. Do you agree with him? Let us know in the comments!
References: