Anyone Can Clone Your Voice

Was this essay forwarded to you? Join for Free Here

In the early 2000s, Rudy Kurniawan, better known as Dr. Conti, made a name for himself procuring and selling fine wine. And we are not talking about bottles worth a few hundred bucks. Kurniawan regularly sold bottles ranging from $2,000 up to nearly $50,000. He was a master at his craft. But his craft was not being a sommelier. He was a counterfeiter.

Kurniawan spent years blending cheap wine in his California kitchen to match the aroma and tasting notes of some of the finest wine in the world. He recreated ancient labels, repackaged bottles with authentically repurposed cork, and emulated the wax seals. He fooled some of the best. As a trained sommelier myself, I find his attention to detail impressive. His blends were nearly indistinguishable forgeries. The FBI estimated he sold around $30 million in counterfeit wine, and by one widely cited estimate, as much as $550 million of his fakes may still be in circulation (Teague, 2018). In March of 2012, agents raided his home, and he was later convicted of fraud (U.S. Attorney's Office, SDNY, 2013). What caught him was a label. He had put a vintage on a bottle that the estate had never produced.

That was only fourteen years ago, and it required a decade of obsessive skill to be able to pull off his forgeries. Today, that level of fraud requires none. And it is potentially far more devastating than some fake wine.

For example, right now, anyone can create a voice clone. In the time it has taken you to read this far, you could have cloned mine. Anyone can create a deepfake from a few seconds of video or a handful of images. Generative AI erased the apprenticeship it used to take to become a forger. Now a free trial account will do. We built the technology far faster than we built the wisdom to hold it, which is the Velocity Gap I keep coming back to.

This is the fourth part of my series on childhood surveillance, and it goes back to the video of my boys and the knocked-over tower. If we place our trust in the proof on the screen, we stop trusting the person standing in front of us. But I ended last week on a question, and it is the one this essay has to answer. What if we can no longer trust the proof either, because the proof can be manipulated too?

Consider a scam that has been around for years. At one of my earlier companies, a controller nearly wired money out the door on an instruction that looked like it came from the CEO. Pulling that off convincingly used to take a sophisticated hacker with intimate knowledge of our company’s inner workings and a few other nuanced details.

Today, an eight-year-old with a few prompts and a cloned voice could do the same thing. And we are seeing it. People are receiving calls from loved ones asking for personal information or for money when really it is a generated voice clone.

What stopped that phishing attempt at that company was not a better tool or system that detected threats. It was a process that made two people get up, walk across the office, and look at each other before the money moved. That human interaction is much harder to fake than an email, a phone call, or even a face on a screen.

For most of history, courts relied on an eyewitness. And every so often there was a false witness, someone paid to say what was not true. It happens on television dramas all the time, and it has happened in real life. Knowing the limits of human honesty, we turned to science for proof that we hoped could not be altered.

We wanted something harder to argue with, something empirical. Take the fingerprint. A certain number of points on a print have to match before an examiner will call it a match. The old convention held that 12 unique points were enough. England uses 16, France 17, and in the United States there is no standard at all. It is left to the judgment of the individual examiner (National Research Council, 2009). One person decides whether a print is a match.

With those human limitations on fingerprint analysis, courts looked for something less open to interpretation. In 1984, Sir Alec Jeffreys discovered genetic fingerprinting, what we now call DNA testing. DNA became the single source of truth. It was irrefutable. We went back and retested old crime scenes to confirm the right people were behind bars, and release those who were wrongly accused. Since 1989, more than 600 people have walked out of prison because DNA proved they did not do it (National Registry of Exonerations, 2025). DNA is one of the more solid things we have. But identical twins share it, and labs make mistakes. It’s closer to perfect, but still not infallible.

Every one of these tools gave us more proof than the last. We have more proof than ever (digital tracking, cameras, smart watches, etc.), and we trust it less than ever.

The polygraph is the clearest case. We built it to be the ultimate source of the truth. As if wiring a person up to a contraption so the body’s signals would give away what the person was hiding. But that failed. The courts still will not take polygraph results, because it was never reliable enough to hang a verdict on (Frye v. United States, 1923).

There is a philosopher named John Gray, an atheist, who described modern unbelief as "a continuation of monotheism by other means" (Gray, 2018). In its place, he argued, we keep building stand-ins for the thing we stopped believing in: humanity, science, technology, one after another.

Go back through the list. The eyewitness. The fingerprint. The polygraph. DNA. The camera. Every innovation was something we reached for so that we would not have to put our faith in each other. Every invention was a surrogate for truth. We tried to trust anything but each other.

This leaves us with two fears, and both are real. A world with no single source of truth is frightening. A world where every person carries a different version of the truth is just as frightening.

Go back to the tower that was knocked over. If we had gone to the camera to review the game tape, there is a solid chance that the proof shown on that camera can’t settle anything. Now, to add onto the fact that footage can be so easily manipualted and faked today, then we have to consider where we put our trust in the first place. It must be earned and exist between humans.

That should reshape the entire surveillance question for a parent. We already worried we were raising an anxious generation, and the deeper worry now is a generation that does not know how to earn trust or to whom truth actually belongs.

It is the same question underneath every conversation I have had on my podcast. How do you trust a kid to ride the subway alone, the way Lenore Skenazy describes? How do you trust them to work it out with other kids with no adult refereeing, the way Kevin Stinehart argues? How do you trust them to carry real responsibility and look after the younger ones, the way Michaeleen Doucleff describes? Every one of those is a question about trust between people, and not one of them is a question a camera can answer.

This matters for leaders both today, with our current workforce, and those on their way to our offices.

The generation raised deepest inside the fake is the one hunting hardest for authenticity and the truth. Gen Z never knew a world without wireless internet, and yet they are the ones filling physical bookstores and helping drive Barnes & Noble to open dozens of new stores. Gen Z are the ones going back to the mall, not to go on a shopping spree but to gather together in-person. Even the often misrepresented return-to-church story, which I have argued is more realignment than revival, points the same direction. The ones who show up are not chasing a trend. They are after something real and authentic.

I have said it before, and it is truer now than ever. For this generation, authenticity is the core indicator of trust. And authenticity is the one thing a feed and an algorithm cannot manufacture.

If Gen Z is the first visible swing back, Gen Alpha is the test of whether it holds. They are as old as thirteen this year, getting their first phones, their first social feeds, and a front-row view of everything generative AI can produce. We are busy teaching them to use the tools. The harder question is whether we are teaching them to sit in a room with another human being.

Just today, I brought my six-year-old into a two-hour meeting. There were eight of us around a table, and by my quick age guessing, four generations in the room. Before we walked in, I told my son to shake their hands, look them in the eyes, and speak clearly when spoken to. It was hard for him. But that red-in-the-cheek, awkward feeling is real, and it is something you can never get from a simulation. It establishes trust in a way that no Zoom call can.

So as leaders, we need to force the uncomfortable. We have to remember those now-nostalgic moments around the water cooler, or walking over to the next cubicle to ask a question, are not just wasted productivity and time theft; they were earning trust that cannot be bought. When we are in person, having real dialogue, it is hard to generate something artificial. Sure, people will be disingenuous. I am not naive enough to argue that if we go back in person, it will be like Mayberry where everyone tells the truth and is an honest human. But, in those human-to-human encounters, if someone is deceitful, at least lying to your face, and that is far better than leaning on technology to create and deceive and deepfake and scheme and defraud.

There was a time when a handshake meant something, when a person's word was their word. We relocated that trust into systems instead, the legal system, the government, science, medicine, the camera, hoping each one would carry it for us. Verifying all of it ourselves is exhausting. We have more information and knowledge and even insight than ever before, but that is not a replacement for wisdom. This is the very friction we thought technology would remove, and instead we have used technology to remove the humanness of it all.

Next week, we will conclude this five-part series examining how the pendulum swings. And it will swing hardest for the ones who never knew a world before the feed.