Fighting corruption is one of the most time-consuming and difficult challenges for modern public and private institutions. Application of the AI tools in this process can be a promising solution, however, first, we need to assess both benefits and risks.​

​

Using AI against corruption can be beneficial due to its tremendous capabilities: AI can process and analyze big data, detect complicated patterns and schemes of corruption fights, it is an autonomous decision-maker and cannot be bribed or coerced, therefore it can assist corruption fighters with investigations and predictions.

​

However, if not implemented accurately, AI can generate biased outputs, make decisions that cannot be fully explained and justified and create a climate of surveillance in the organization. Also, AI tools should have a very high level of accuracy, as reputational losses associated with mistakes can be extremely high. Last but not least, AI tools, if not included in the systematic approach, cannot fix the problem of corruption alone.

​

To mitigate these risks, we recommend AI judgments be further examined by human analysts, outputs are accompanied by clear justifications and descriptions of decision-making process flow and collected data be digitized, checked, cleaned, and updated regularly; Application of the AI tools should be carried out with respect to human rights.