How do you know you’re talking to a human?
Last month, a story went viral about a man receiving the phone call every husband dreads: “I’m at the emergency room.” His wife explained that their son had fallen off his bike and gotten hurt so bad he needed emergency medical care. And to make matters worse, the hospital wouldn’t take their insurance. The wife needed cash ASAP.
The husband asked an odd follow-up question: “What’s our passphrase?”
The wife was aghast. She kept repeating the request for him to wire money. But he insisted. Finally, he hung up.
Then he called back his real wife.
He was savvy enough to know that his “wife” who called him was an AI deepfake, using a cloned phone number. But the one thing the fake wife didn’t have was their family passphrase: a secret never kept online and never even written down. It was a code word passed between them in a way the surveillance overlords weren’t able to record.
Oddly enough, the same day I read this story, I came home from a trip to a similar situation, but in reverse. My 9-year-old son looked at me in a strange way the whole time I was settling in, and finally asked, “What’s your favorite color.” I answered to his apparent satisfaction, but he gave me a few other quizzes.
Finally, he sat in my lap and broke down crying, confessing he needed to tell me something. And he kept saying, “Don’t be offended.” Then he said he was worried I was an impersonator. A robot disguised as his dad. A deepfake.
Now, he’s never seen Terminator, but he understands the concept well enough. It’s haunted his nightmares. After all, he’s living in a cyberpunk world that didn’t exist when I was his age.
The next morning at breakfast, I read my kids the story of the man with the fake wife at the ER. We talked about how we should also have a family password. Just in case one of us ever gets a spoofed phone call with an AI voice clone. And how we need to discuss our password somewhere with no devices possibly spying on us.
It’s a fun thing to prepare for, even if it’s unlikely.
This all leads me to another question: in a world where robots can make phone calls, create music, take photos, and write stories, how do we know we’re dealing with humans?
The common wisdom used to be that an email full of typos meant it was a scam, but nowadays scammers use AI, which never makes a spelling mistake. So it seems the true way to prove your a human is to intentionally mistype a word. Maybe even a very specific word.
Beyond these important security measures, I think the bigger issue is that so much of culture is now being built by robots. So much of online media is made with AI tools, and I’m starting to think even the accounts posting it are themselves automated. We even have military conflicts being decided, reported and analyzed by AI, all while propaganda in every direction is created by AI.
It seems more important than ever to establish what makes humanity important.
How do you prove you’re a human?