It was the early 2000s. My first startup was finally getting some traction, which meant I was working 100-hour weeks and living on caffeine and adrenaline. Then I get a call. It’s my bank. They wanted to know if I’d just spent two grand on live lobsters.
I had not.
That was my first taste of credit card fraud. Back then, finding thieves was a joke. It was a bunch of analysts in a back office, staring at spreadsheets, trying to spot something fishy. It was slow, clunky, and the bad guys were always miles ahead.
Today? It’s a whole different world. Every single time you use your card, there’s an AI standing guard. It’s a war fought in the milliseconds between a swipe and an approval, and AI is the only reason we’re winning. As an investor in over 200 companies, including AI pioneers like Anthropic and OpenAI, I’ve seen this transformation up close. Machine learning isn't just making fraud detection better; it's changing the entire game.
A Problem Measured in Billions
Let's be real, the numbers are insane. In 2024, Americans lost over $12.5 billion to fraud. That's a 25% jump in just one year. Credit card fraud is the most common flavor of identity theft, hitting almost half a million people. These aren't just numbers on a page. That's half a million people whose lives just got a lot more complicated.
And for the banks? It's a nightmare. It's not just the money they lose on the fraud itself. It's the cost of investigations, issuing new cards, and the hit to their reputation when they fail to protect their customers. It's a multi-billion dollar problem that's not going away.
How AI Fights Back
So what’s the secret weapon? It’s all about patterns. Old-school fraud detection used simple rules. A transaction from a weird country or for a huge amount would raise a flag. But criminals aren’t stupid. They adapted, making smaller, more normal-looking purchases to fly under the radar.
AI doesn’t play by those rules. It learns. We feed it billions of transactions, and it starts to understand what your normal looks like. It spots tiny details and connections a human could never see.
Here’s how it works:
Real-Time Muscle: This is the core of it. AI analyzes transactions as they happen. In that split second, it’s looking at hundreds of things—the time, the location, the amount, your spending history, even the device you're using. If it smells a rat, the transaction is dead before it starts.
You Are the Password: This part is just cool. AI can learn your behavior. How fast you type, the angle you hold your phone, the way you move your mouse—it all creates a digital fingerprint. So even if a thief has your password, they can't fake being you. The AI will know.
Connecting the Dots: Fraudsters are rarely lone wolves. They operate in networks, using the same devices or shipping addresses. AI is brilliant at finding these hidden connections and taking down entire fraud rings in one go.
Fighting Ghosts: Synthetic identity fraud is a huge problem. Criminals invent a person from scratch using a mix of real and fake data, making them incredibly hard to spot. But AI can see the cracks. It notices the weird inconsistencies, like a brand-new credit file suddenly applying for ten different loans.
An Investor's Take
I get pitched by startups all the time. When it comes to AI and fraud, the companies that get my attention are the ones thinking bigger. It's not just about blocking a bad transaction. It's about predicting fraud before it even happens. They're building systems that find the weak spots in a company's defenses and fix them.
And now we have generative AI in the mix. To catch a thief, you have to think like one. We can now use generative AI to create fake fraud attempts—millions of them—to train our detection models. It’s like a sparring partner that never gets tired, making our defenses stronger every day.
The Game Never Ends
But let's not get complacent. The bad guys are using AI too. They're creating deepfake videos, smarter phishing attacks, and all sorts of new scams. This is a cat-and-mouse game that will never, ever end.
Winning means staying paranoid. It means pouring money into R&D, getting the government and private companies to actually talk to each other, and making sure everyone knows this is a real threat. It also means we all have to be smarter about our own data.
I’ve come a long way from being the guy who accidentally bought a boatload of lobsters. I’ve seen what AI can do, and I’m genuinely excited about how it can protect us. It’s not a perfect solution, but it’s the best shot we have. And I’m all in on that fight.
Frequently Asked Questions
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.