Caught most of Catch Me If You Can on TV a few weeks ago. Had not watched it in years. Twenty-plus years after the movie came out, it reads differently.
The movie tells the story of Frank Abagnale Jr., a teenage con man who, throughout the 1960s, impersonated a PanAm pilot, a Georgia doctor, a Louisiana attorney, and a Brigham Young University sociology professor. He cashed millions of dollars in fraudulent checks before the FBI finally caught him. Steven Spielberg made the movie. Leonardo DiCaprio played Abagnale. Tom Hanks played the FBI agent who spent years chasing him.
The reason the story still works as cinema is that being Frank Abagnale was hard. He had to physically obtain a PanAm uniform. He had to study check-printing techniques. He had to memorize medical terminology. He had to land planes mentally, charm bank tellers, learn the cadence of a courtroom. Each forgery cost him hours of effort, considerable risk, and an unsustainable amount of attention from people whose job was to catch him.
That world is gone.
For most of human history, being honest was the cheap default. Lying took effort. The math worked because pulling off an Abagnale required an Abagnale.
For most of human history, being honest was the cheap default. Lying took effort. Forgery took resources. Impersonating someone took time and craft. The entire infrastructure of trust we built around modern commerce, the signatures and witnesses and seals and identity documents, was designed to make dishonesty expensive enough that most people would not bother. The math worked because pulling off an Abagnale required an Abagnale.
Then quietly, in the last few years, that equilibrium broke.
AI has separated the cost of producing a lie from the cost of detecting it. The teenage version of Frank Abagnale today does not need to find a PanAm uniform. He does not need to memorize legal procedure. He does not need to charm anyone in person. He needs a twenty-dollar monthly subscription and twenty minutes of patience. The lie is cheap. The truth is expensive.
AI has separated the cost of producing a lie from the cost of detecting it.
There is a particular kind of irony here. We spent two centuries building a civilization on the assumption that lying was hard. Notarized documents. Witnessed signatures. Photo IDs. Voiceprints. The whole edifice of modern commerce was a quiet bet that being dishonest would always cost slightly more than being honest. The bet held for a long time. It held even when Abagnale was running checks across the country, because what made him exceptional was precisely how rare and difficult what he was doing actually was. Most people were not going to try, because most people were not Frank Abagnale.
Then we built a machine that lies for free.
The headline number, the one that has actually changed, is that honesty itself is no longer free. And almost nobody in a board meeting is allowed to say so out loud.
What we did not also build was the institutional vocabulary for talking about what we did. Banks still talk about fraud losses. Regulators still talk about KYC compliance. The conversations are about the rounding errors. The headline number, the one that has actually changed, is that honesty itself is no longer free. And almost nobody in a board meeting is allowed to say so out loud.
The Abagnale-era infrastructure of trust assumed that the friction of being dishonest would protect the system most of the time. That assumption was load-bearing. The frame and beams of our institutions rest on it. And we just changed the assumption without telling the building.
That is the structural shift behind the fraud headlines.
Honesty got expensive. Dishonesty got cheap. We have not yet started designing for what happens next.
Shyam Menon is a product leader specializing in fraud and identity in financial services. This is part of an ongoing series on the architecture of trust in financial services. He writes at shyammenon.com.