The History of Lie Detectors—And Why They Don’t Actually Work

At some point, we all dream of a world where lies are instantly, dramatically exposed. Like, you tell your boss you’re “almost there,” and a buzzer sound announces, “THAT was a lie.” The idea of a machine that can cut through lies has been a kind of holy grail for law enforcement, suspicious partners, and reality TV producers for over a century.

Enter the polygraph, better known as the lie detector—a device that promises to decode your secrets using little more than a blood pressure cuff, a few electrodes, and the premise that your body is a terrible liar, even if you’re not. It looks scientific, sounds impressive, and it’s been a fixture in pop culture for decades. But here’s the catch: it doesn’t actually work. At least, not the way people think it does.

So how did we get here? And why are we still here? Let’s rewind.

Ancient “Science”

Before the polygraph, before the beeping wires and official-looking charts, humans were already experimenting with ways to catch liars. Spoiler: most of those ways were either absurd, brutal, or both.

In ancient China, one early “lie detection” technique involved forcing suspects to chew dry rice and spit it out. The theory went like this: liars get nervous, nervous people have dry mouths, and dry rice plus dry mouth equals suspiciously unsoggy evidence. No rice pudding? You’re guilty.

Meanwhile, in parts of medieval Europe, lie detection was outsourced to God via “trial by ordeal.” Accused of something? Cool, just grab this red-hot iron bar or plunge your hand into boiling water. If your wounds heal nicely, congratulations—you were innocent all along. If not, well… at least your sins are now known and lightly charred.

The Polygraph Is Born

The modern lie detector got its big debut in 1921, when a California police officer and part-time medical student named John Augustus Larson invented a device that could track a suspect’s physiological responses—specifically blood pressure, respiration rate, and pulse—while being questioned. He believed that lying triggered stress, and stress left a measurable trail.

His machine wasn’t perfect (and honestly, wasn’t even great), but it was the first to bring a clinical aesthetic to the ancient art of “guilty or not guilty?” And people loved it. Finally, a way to put nerves under the microscope!

By the 1930s, the polygraph had gained traction with law enforcement. The FBI, led by the drama-loving J. Edgar Hoover, latched onto it as a symbol of modern policing. For the public, the image of a twitchy suspect hooked up to wires became shorthand for serious interrogation.

But the whole thing rested on a pretty fragile assumption: that guilt always looks like stress, and that stress always means guilt. And as it turns out, that’s just… not how people work.

What It Actually Measures: Not Lies, Just Stress

Despite its name, the polygraph doesn’t detect lies. It detects changes in your body like heart rate, sweating, breathing, and blood pressure, that might indicate you’re feeling anxious. It’s basically a glorified stress thermometer. A moody Fitbit with a clipboard and a grudge.

Sure, lying can make you nervous. But so can being accused of something you didn’t do. So can trying not to cry in front of a stern examiner. So can the caffeine you pounded on the way to the testing center. Or the fact that you’re currently strapped to wires and being asked, “Have you ever stolen from an employer?” even though you once took a Sharpie home, and now you’re spiraling.

Worse, not everyone responds to stress the same way. Some people—sociopaths, spies, poker players, teenagers—can lie with all the emotional ripple of a houseplant. Others are just naturally anxious and might fail a polygraph while trying to remember their zip code.

The Data (or Lack Thereof)

You’d think that a device used in national security screenings and high-stakes criminal cases would have some serious science behind it. Nope.

The National Academy of Sciences, the American Psychological Association, and pretty much every major scientific body that’s looked into it have all said some version of: this is not a reliable way to detect deception. Multiple studies have shown the polygraph’s accuracy to be somewhere between 60% and 80%. Which sounds okay—until you realize that flipping a coin is 50%, and the polygraph has ruined a lot more lives.

In one infamous case, a man named Gary F. falsely confessed to murder after failing a polygraph. Why? Because he panicked. And the machine interpreted his panic as guilt. Turns out, he was innocent. The real killer was caught later.

And it’s not just false positives. Polygraphs are also comically easy to beat. There are entire YouTube tutorials and Reddit threads dedicated to “countermeasures”—ways to manipulate your breathing or subtly cause pain during control questions to throw off the baseline. Don’t worry, I won’t link them here. But they exist. And they work often enough that even the CIA admits it’s a problem.

So Why Are We Still Using Them?

Two words: psychological leverage.The polygraph survives because it looks official. It sounds scientific. And when you’re hooked up to a machine and someone says, “Tell the truth or the test will catch you,” it can be really hard not to crumble—even if you’ve done nothing wrong.

Law enforcement still uses polygraphs in the U.S. during investigations, especially to rattle suspects or coax confessions. Some government agencies (including the CIA, NSA, and certain police departments) still require them for pre-employment screenings. Which is wild, because you can literally pass one while lying or fail one while telling the truth.

They’re rarely admissible in court—most judges treat polygraph results like ghost stories. But in the court of public opinion? In job screenings, parole hearings, and custody battles? The machine still holds a dangerous amount of sway.

What About New Tech?

People keep trying to invent better lie detectors. Functional MRI (fMRI) scans, eye-tracking software, thermal imaging, voice stress analysis, brainwave tech. All promising, all still deeply flawed. At best, we’re still measuring things around the truth—emotions, tension, arousal—not truth itself.

Because here’s the frustrating reality: lies don’t have a fingerprint. There’s no universal “I’m fibbing” signal that shows up on a machine. Lying is a complex, psychological act that depends on context, culture, and conscience. It’s not something you can fully quantify with squiggly lines and a clipboard.