What's the connection between being an AI engineer and being a detective?
Does the AI engineer also work incognito? Is he, like the detective, carrying a grey powder and trying to blend in? Or is he eavesdropping or gathering evidence on a case? Well, the answer to that question is much less obvious than it may seem. It turns out that the system engineer is connected to the detective money laundering, namely-the prevention of this process.
From this article you will learn how artificial intelligence (AI) supports the work of bank analysts who track illegal transactions. We’ll also tell you what you can do to become a digital detective in the future.
What do Al Capone and AI have in common?
In the 1920s, the United States government banned the consumption of alcoholic beverages. What did the rulers want to achieve? Certainly reduce the abuse of alcohol by Americans, as well as counteract the phenomenon of hunger (caused by a devastating war) by increasing the share of grain in bread production; previously, a lot of agricultural fruits were “wasted” on alcohol. Residents of the United States were not satisfied with the restrictions and were looking for ways to circumvent them.
This situation has obviously been exploited by the criminal community, m.in. the famous Al Capone, who was at the head of the mafia operating in Chicago. Criminal organizations profited enormously from the illicit trade in alcohol. Wanting to put the proceeds into legal circulation, Al Capone opened a chain of laundries – of course, legally. By illegally inflating the price of each of his units, he was gradually putting the money earned from the alcohol trade on the market. This is where the term “money laundering” comes from, which we still use today.
While the ban on alcohol consumption was quickly withdrawn, the problem of money laundering remained topical. Governments and international law enforcement organisations continue to take radical steps to keep illegal activities to a minimum. In many countries, strict rules have been introduced requiring financial institutions, such as banks, to monitor cash flows and inform the relevant authorities of suspected money laundering or terrorist financing.
Banks have no choice-the penalties imposed by the supervisory authorities are huge. Financial institutions that hide information about money laundering face not only financial consequences, but also legal ones-including the loss of their operating licenses. Banks have been obliged to introduce systems to monitor the flow of money, such as regulatory systems. They also opened internal AML units (anti-money Laundering), and it is precisely the employees of these units – AML analysts – that we can safely call detectives. Why?
It turns out that people who are engaged in combating money laundering in banks act as detectives-on the basis of available evidence (photos, reports on the internet, information from special units etc.) can determine whether they are dealing with money laundering or not.
How can you launder money today?
The money laundering strategy that Al Capone invented 100 years ago is still relevant today, although criminals, for fear of detection, are using increasingly complex techniques. They can, for example, set up a network of different companies between which fake invoices will be sent, Play Poker at a set table in a casino… And who says you can’t offer a lottery winner more than his ticket is worth? Thus, the criminal will be able to boast of “great happiness”, and the one who sold him the coupon will have to explain to the police how he got several million in cash from a man who, as it turned out, used a false identity. There are a lot of ways.
On the other side of the barricade
Just as criminals have their methods of putting illegal money into the market, so banks have their ways to detect it-they use this to do so m.in. previously mentioned rule systems.
The bank’s AML process begins with an analysis of all transactions carried out by the institution’s customers in order to detect any suspicious behavior. Suspicious behavior is defined by rules developed by AML experts. These rules trigger alerts that indicate which of the transactions made by a bank customer may have signs of money laundering. The AML analyst takes one of these alerts under the magnifying glass and carefully checks the transaction – who is the sender, who is the recipient, looks for information about the activities carried out by these persons, checks the databases. The analyst combines all the individual information into a whole and produces a report that determines whether the client is actually laundering money.
The specificity of simple rule-based systems means that in addition to real money laundering cases, legal transactions can also be labeled as suspicious. There are thousands of such false alarms.
Banks and other financial institutions are keen to use new technologies to streamline their internal processes. Increasingly, AML analysts are supported by tools that use artificial intelligence solutions. In this area, our digital detectives-artificial intelligence engineers-can prove themselves.
Shoulder to shoulder
Supporting the work of an AML analyst with tools enriched with AI solutions is an example of the ideal combining human work with new technologies. New technologies, behind which stands a second man-a machine learning engineer.
Okay, but how do you actually know which transactions are the most suspicious? Criminals aren’t stupid. Their transactions are unlikely to be distinguished by anything special: they will not transfer large sums, the target countries will not necessarily be tax havens. Hundreds of thousands of transactions for different amounts, from different countries can certainly make you dizzy, and the answer to the questions of where to start and where the illegal transaction can hide, seems unattainable. In this situation, our analyst AML comes to the rescue digital detective-machine learning engineer, which builds tailor-made solutions based on artificial intelligence (machine learning, deep learning). Its task is to analyze and combine huge data sets.
The result of the work of a machine learning engineer can be m.in. build an alert ranking, that is, an indication of which alerts are more risky and should be looked at first and spend more time. It is much easier for analysts to work when they receive additional information about which transactions may be real cases of money laundering. By analyzing these transactions first, information about the suspected crime will reach the relevant units much faster.
The work of an AML analyst is also tedious collecting traces, most often online. Repeatedly need to visit facebook, twitter, internet maps, government databases, as well as search the websites of companies. You need to answer yourself a lot of questions, for example, whether the way you spend money makes any sense and is consistent with the client’s profile. To connect the facts and connect the details, you first need to get the right information. And here again our “digital detective” comes to the rescue! Thanks to the modern tools for automatic data collection and analysis built by him and other developers, analysts receive support in the process of getting to know the bank’s customer. The evaluation developed by them contains information that is crucial from the point of view of our “investigation”.
Tools for visualizing financial flows, combined with geolocation, also help to better understand the specifics of the case. The digital detective is able to use artificial intelligence to process and draw conclusions from huge amounts of data that would be impossible to analyze in the traditional way.
Al Capone, who created and initiated the money laundering procedure 100 years ago, certainly did not expect that specialized AI models would stand guard over the rule of law. Well, he probably didn’t even know that something like artificial intelligence would ever come up!
Become our digital detective
Are you looking for an interesting way to spend your next vacation? Do you want to be part of a team that creates completely new solutions? If you like to look for dependencies, support your teammates, and know the basics of programming, then feel free to apply for the Comarch internship program.