Crime doesn’t pay — any TV cop will tell you that much. But, what about their counterparts — those slick TV criminals that, season after season, manage to elude the law with clever schemes, quicker wits, and just plain dumb luck?
What about Walter White, the science teacher turned meth-making mastermind in Breaking Bad, who had so much money that at one point his wife had to rent a storage unit just to store a literal mountain of it? Or how about Marty Byrde, the anti-hero of the hit Netflix series OZARK, who goes from white-collar financial scams to laundering millions of dollars for Mexican cartels in just a few minutes of TV?
Our anti-heroes are great at evading the law, but what about the banks? Walter White’s schemes may have eluded financial services back in 2012, but today’s banks are stocked with AI and machine-learning-powered risk assessment and anomaly detection models that scan millions of transactions per second. Marty Byrde, for all his talk of offshore havens, may also have a tougher time today, as international banking systems embrace ML-powered anti-money laundering systems.
The question, then, is how long could these two TV masterminds hide their ill-gotten fortunes from financial institutions with all the technology available to detect financial crime today? Let’s take a look.
(SPOILER ALERT: we’re discussing all seasons of both shows, so you’ve been warned.)
Without getting too far into it, we’re focusing on both Walter’s and Marty Byrde’s laundering operations, which work slightly differently, but seem to have similar results. Also, it’s worth noting that while Walter’s money laundering plays a slightly smaller role in Breaking Bad, OZARK is centered around it. With those caveats in mind, let’s dig in.
For Walter White, money laundering becomes a necessity due to the resounding success of his meth-making and selling. He has too much cash, and he literally doesn’t know what to do with it, so he buys a carwash. How would that work, exactly? Let Saul Goodman, lawyer to the Breaking Bad crime world, explain:
So it’s easy as one-two-three: placing the money into someone’s hand, layering it into a cash-flow business’ books, and integrating it into its revenues. Pretty simple, right? What about Marty Byrde, though? He probably has some complex scheme. Once again, take it away Marty:
Essentially, Marty and Walter operate similar laundering schemes. The idea is to filter in ill-gotten gains with clean cash by cooking the books of cash-heavy businesses and moving that newly cleaned money into the bank. This works great in TV shows where the banks are plot devices, but might be less so in the real world, where banks can’t afford to let money laundering simply happen.
What would happen if Marty and Walter were attempting to launder their millions in the real world, where AI and ML-powered detection tools have become the norm? A sudden influx of cash into an account that had been relatively stable is an indicator of suspicious behavior for most machine learning systems, and would at least mark an account for observation by banks. Even so, this may be explainable by a large lottery win, or, as Walter claimed, a successful gambling outing.
Similarly, Marty’s first suspicious activity was withdrawing a massive amount of money from one bank account ($8 million) and moving it elsewhere. Again, though, these kinds of things DO happen sometimes, but such a large sum would undoubtedly be pinged by an AI-based anti-money-laundering (AML) detection system, which can track millions of transactions per second to find even the smallest anomalies that could indicate a crime.
It gets even dicier when they actually start layering their money into cash businesses and creating even a small paper trail. Most ML-powered anomaly detection systems aren’t looking for big smoking guns, but examining millions of small transactions to find unusual patterns and behaviors.
A chemistry teacher with a history of mounting medical costs suddenly making hundreds of thousands of dollars at a carwash they just bought with a large windfall would be a pretty suspicious pattern. Moreover, Walter couldn’t even filter in that much cash without raising serious eyebrows — banks and financial institutions generally know the rough numbers a business might put up monthly, so unusual numbers start ringing ML-powered systems’ alarm bells.
Marty’s schemes would also raise eyebrows based not just on historical data, but on comparative statistics. While it’s not unusual for hotels to purchase equipment and make repairs and renovations, Marty’s operation revolves around over-ordering, overpaying, and buying more expensive options to justify the large cash flows. ML-powered systems that incorporate external data could pick up on a few anomalies in seconds. The first is that by measuring the size, location, and relative performance of competitors to the hotel, banks could establish a suspicious pattern of purchases and expenditures. Marty’s problems would only get worse once he starts moving a small hotel or strip club’s finances into an offshore bank, raising eyebrows about the type of money being passed through.
Sudden spikes in revenues aren’t necessarily fishy, nor are raised expenditures, but if the numbers are high enough, they’re a massive red flag for banks. This would mean that either Walter and Marty need to be okay with playing the (extremely) long game, or risk being spotted by AML systems in the banks they’re trying to filter their money through. Banks, who are increasingly under pressure to comply with stringent AML regulations and using much more sophisticated technology, could start detecting these irregularities quite easily, and quite quickly.
It’s true that money laundering still happens in large numbers across the world, though the schemes in use are incredibly sophisticated and complex. Unfortunately for Walter and Marty, their operations don’t quite reach that level. The reality today is that most banks and financial service providers are under incredible pressure to comply with AML regulations, and have embraced machine learning and AI as powerful tools in that endeavor.
Even the most seemingly harmless behaviors would immediately be marked as suspicious for evaluation, and lead to greater scrutiny from banks. This is the power of machine learning in the real world. While TV banks rely on an ace investigator to piece the puzzle together, real banks have technology that makes basic schemes harder to pull off because of the sheer analytical capability they possess. A system that can scan a million transactions in a few seconds and uses data from a variety of sources can mark even the smallest anomalies quite fast.
While they’re terrific TV viewing and masterminds in their own right, it’s tough to see how Marty and Walter could go on indefinitely hiding their money from banks, financial institutions, and tax authorities for that long.