People have more to offer than machines. We hear machines will transform all aspects of our lives, including how AML is done.

Reading AML software, consulting, and outsourcing firm marketing, it seems the equation goes something like this:

algorithms + data + machines + false positives + AI + Machine Learning = efficiency = less work = fewer people = less cost.

At this point, I lose track and am unsure where uncovering more actual crime, money laundering, and terror financing comes in.

As a human myself, I want to speak up for us homo sapiens sapiens. We’ve been around in our current form for about 70,000 years. During that time, we made some good progress (language, medicine, air travel) and had some setbacks (war, famine, boy bands).

The connection between humans and strong AML compliance is clear. The purpose of AML work is to identify, investigate, and report suspicious activity. Doing this well requires a few things humans are good at. Among them are:

Critical Thinking – Humans are uniquely able to think critically and solve complicated problems. The software provides data and tools, but people interpret data and draw conclusions. Experienced investigators have seen examples of how, in one context, transactions are suspicious, and in other contexts, they are not. Think about international wire transfers to high-risk countries. On their face, they appear concerned, and in many cases, they are. Why is a company, recently incorporated in a secrecy haven, sending large wires to Syria? However, what if it’s a long-operating, unblemished NGO sending funds to a relief agency that, after research, is also validated to be legitimate with no reputation concerns?

Quality, not just speed – Software, particularly in AML, is seen as a way to move things fast. Sometimes, fast is not good. A good investigation may take time, thinking, and bouncing ideas off co-workers. It is important for machines to accelerate the gathering of data and presentation of it to investigators for analysis. But just because the machine speeds up getting data, investigators often need to move slowly to think through issues, identify new areas to research and write out their observations and rationale.

New, Novel situations – Humans are better at spotting something new or ambiguous. Even AI requires it to be trained on existing data. True, most money laundering techniques are old, but new variations arise frequently. Spotting these new variations is what separates good investigators from the average. Criminals are constantly mixing things up for the very purpose of confusing investigators. Endless variations of laundering appear and will continue to. It takes an experienced human to spot the new approach and dig deeper.

Creativity – We hear from law enforcement that to get attention, make the SAR narrative tell a story. ChatGPT is dull. Now, Grok, Elon’s new AI, maybe a different story, but until then, human writing remains better.

I realize I am biased. I like humans. I’m on their side. And for good AML investigations, I still take a human over a machine.

Back to Top