The researchers found that using artificial intelligence to predict future corruption scandals, criminal schemes and prevent them.
According to the world Bank, every year corruption schemes “deflate” from the global economy over $2 trillion.
Thus, corruption has become one of the main causes of slow economic growth and socio-economic inequality. A new study provides the key to solving this problem.
So, scientists led by Matthias Pepper of the University of Maribor (Slovenia) and Haroldo Ribeiro from the University of maringá (Brazil) with the support of the Vienna scientific centre published a paper on his analysis of political corruption, using AI algorithms.
In their work they have built a dynamic network of political scandals that have occurred in Brazil over the past 27 years, and analyzed how the corruption network has evolved over this time. The model includes more than 400 points representing people and the links that connect participants more than 65 well-known corruption scandals. As it turned out, the neural network is able to reveal all the details of corruption scandals.
“Despite the veil of secrecy that surrounds corruption, we show that the application of the methods of network science reveals the essence of politically corrupt conduct. Science can expose people who are doing everything possible to remain anonymous,” said Matthias PERC, commenting on the results of the study.
Scientists have found that the algorithms are capable to identify people who played a Central role in several corruption schemes. In addition, it was found that political scandals usually coincide with the election cycles, and the corruption involved small groups of people (maximum eight people), because it is easier for them to zakonservirovat.
“We have found that corruption is reduced to the selfish behavior of small groups in hierarchical networks business. We also found that only a few people dominate the modular structure of the network that often suddenly changes as the change of government, and, ultimately, we show that future “partners in crime” can be accurately predicted based on the dynamic structure of corrupt networks. Thus, we show that the politically corrupt behavior reveals almost all the secrets in the analysis of algorithms,” – said Matthias PERC.
In the future, scientists intend to apply these methods to the problem of corruption in research funding (especially in Slovenia).
“We plan to explore a wide range of problems: from research funded by the government, to a national system of grants. Now they are often directed to support obviously worthless research conducted with loyal people and institutions. These operations do not support talented young scientists. We hope that our research will help with the transition to a sustainable and more equitable expenditure of public research funds,” said Pepper.