Could AI Be Used to Stop Corruption?

An intriguing blog from Ukrainian IT business Sciforce argues that artificial intelligence could be used not only to identify fraud, but to fight corruption too.

To curb bribery and corruption… companies are seeking out new technology solutions to manage what was often relegated as an employee management issue.

They argue that new developments make it feasible to use automation to target corruption.

The next step in protecting against financial crime would be to leverage advancements in artificial intelligence (AI), and data analytics hoping that technology-assisted data analysis can provide the diligence and reliable quality control needed to provide governments and businesses with conclusions that can be trusted.

Some of the AI techniques used to identify financial crime risks include advanced entity resolution and verification, Ultimate Beneficial Owner analysis, deep Web analytics, NLP (Natural Language Processing) Web scraping, network analysis, and volumes and values analysis.

Sciforce emphasizes the use of technology to detect anomalies in data, a concept also expounded by Subex when they launched Crunchmetrics, a brand that focuses on anomaly detection across different sectors. The Sciforce article argues that machine learning (ML) is better at identifying patterns than employing people to do the same job. As a consequence, ML will spot the exceptions indicative of crime.

A surveillance ML-driven tool receives historical information and learns to recognize acceptable and appropriate transactional patterns and then have the ability to identify transactions that do not “fit” that pattern, falling outside of the normal flow of business, and may be anomalous, such as ill-timed or duplicative payments, falsified invoices, and other suspicious transactions.

But are they neglecting the likelihood of false positives? Adrian Harris of Xintec is an expert in machine learning and fraud management, and last year he wrote a Commsrisk article that discussed the advantages of automated detection of fraud whilst being careful not to exaggerate the power of machine learning.

Searching for something in a hay stack when you don’t know you’re looking for a needle takes a lot of time and effort. You will find the needle in the end and you might find some other interesting things along the way, but it isn’t as easy as you expect.

The Sciforce analysis also begs the question of how machine learning will spot a pattern when every transaction is corrupt. To illustrate how this can happen, consider that Ericsson recently agreed to pay USD1bn as punishment for years of corruption, some of which involved creating new corporate entities whose sole purpose was to funnel bribes to government officials and telecoms executives.

Despite the possible shortcomings, it is encouraging to see technologists striving to implement new solutions for an age-old problem like corruption. You can read the Sciforce blog here.

Eric Priezkalns
Eric Priezkalns
Eric is the Editor of Commsrisk. Look here for more about the history of Commsrisk and the role played by Eric.

Eric is also the Chief Executive of the Risk & Assurance Group (RAG), a global association of professionals working in risk management and business assurance for communications providers.

Previously Eric was Director of Risk Management for Qatar Telecom and he has worked with Cable & Wireless, T‑Mobile, Sky, Worldcom and other telcos. He was lead author of Revenue Assurance: Expert Opinions for Communications Providers, published by CRC Press. He is a qualified chartered accountant, with degrees in information systems, and in mathematics and philosophy.