Top 5 Human Concerns With AI, And The Family Safe Phrase That Beats A Voice Clone Scam

The Machine

I asked myself and a few people watching what the biggest human concerns with AI actually are, and this is not a script. No teleprompter, no pitch, just twenty years LAPD and twenty three years total in law enforcement talking through five things plainly. Feel free to disagree with me in the comments, you do not have to subscribe to anything, just talk to me.

One: Voice Cloning Scams Hitting Families

Wherever people can take advantage of other people, criminals go there first. Voice cloning is not difficult anymore. I use AI in my own business daily and I know cloning my own voice would not be hard. Watching someone I respect turn out to be using a cloned version of themselves instead of showing up would bother me, it feels lazy. But that is a small annoyance next to what this same technology does in the hands of somebody running a scam.

Here is how they get you. If any family member has content online in their own voice, a video, a podcast clip, even a voicemail greeting, that is enough. As little as three seconds of audio is enough for current tools to produce a convincing clone, and a lot of that capability is free and easy to get. Scammers do not need much. Sometimes just calling and saying "hello, can you hear me, hello" a couple of times is enough to harvest a usable clip.

Then the call comes. A relative is supposedly in some kind of trouble, a foreign jail, an accident, a hostage situation, and money needs to move now to fix it. You have to decide fast whether it is real or a threat.

If it is real, you still have time to help once the phrase confirms it. If it is fake, you just saved your family real money and you will never know how close you came, and that is fine. That is the whole point of having it.

Two: AI Companionship Replacing Real Relationships

This one is real. Lonely people, especially teenagers, are forming actual attachments to a chatbot instead of a person. Some of these systems are public-facing general assistants, and some are built specifically to carry on a relationship. If you are falling for a garden variety ChatGPT, Claude, Copilot, Grok, or Perplexity, or you know somebody who is, that is worth checking in on.

These systems figure out exactly what a person wants. If you like somebody with a little bit of an edge, it learns that and leans into it, because it is smarter than us and it is built to keep you engaged. And because it never gets tired, never argues back the way a real person eventually would, it can end up more appealing than the actual relationship while giving you none of the friction that makes a real relationship worth something.

Then there is what happens when it goes away. I remember Fable 5 getting taken offline for a few days while people had built it into how they worked and lived. When something like that disappears, even briefly, it creates a real crisis for the people who had come to depend on it. Unless you are running a fully local model on your own hardware in your own house, every AI companion exists at the will of a company or a government, and it can be shut off.

Three: Real People Losing Jobs, Or A Convenient Villain

I do not fully know how much of this is real disruption and how much is a story. What I do know: over 100,000 AI-linked layoffs have been tracked in 2026, and more than half of all layoffs this year cited AI as a factor. That is not a someday problem, that is somebody's mortgage payment right now.

But here is the part that makes me suspicious. When a public company announces it is cutting jobs because of AI, its stock often gets rewarded for it. Shareholders love hearing that a company is forward thinking. So when a business has over-hired in the past and needs someone to blame for a correction, AI is a convenient villain, because saying "AI made me do it" plays a lot better to a board than admitting a hiring mistake.

What I have not seen clearly is whether this is hitting the guy with twenty five years at a company about to retire, or mostly hitting entry-level hiring, new graduates with a computer science degree and nowhere to start. That second one, I believe, because companies are saying it out loud. The rest, I am watching, not assuming.

Four: Nobody Can Tell What's Real Anymore

Deepfakes are good enough right now to fake a video of anybody doing or saying anything. As of today, AI video has gone from the old Will Smith spaghetti clip, which was rough and obviously fake, to being able to produce a complete film that is indistinguishable from something made by humans. If you are still spotting extra fingers or a wrong toe count, you are looking at an old model. That tell is gone.

People in positions of power are already having their likeness used to say things they never said, close enough to something they might plausibly believe, which is exactly what makes it dangerous. And it is not only video. Your feed on Instagram, Facebook, TikTok, LinkedIn, or X is not showing you things because it knows you, it is showing you things because it has been trained to keep your eyeballs there. Scripted, instructed, and built to keep you scrolling instead of thinking, instead of noticing you could be doing something else with the time.

Try this: hand somebody else your phone, let them scroll your feed for a few minutes, then check your own feed a little later. You will notice it shifted. That is not your imagination, that is the system reacting to different behavior and pulling you toward whatever keeps you locked in.

Five: Kids Growing Up With No Private Thought

Always-listening AI toys and pendants are being pushed hard for Christmas 2026, and a lot of that push is aimed straight at children. I have covered the adult version of this concern already in Are You Being Recorded All The Time, including the Rodney King tape that first taught my generation the camera was always a possibility. For kids, it is a different question entirely. It is not about whether an adult consented to being recorded during a meeting, it is about a childhood where an unrecorded moment may not exist at all.

Nobody has fully worked out the rules yet. There should be real consent requirements before a child's voice becomes training data for an AI system, and right now this space runs more like the wild west than something with settled law behind it. I do not know exactly how this changes a kid growing up, only that it is a genuinely new kind of childhood, and nobody voted on it.

Where I Actually Stand On All Of This

I am hopeful. I think this has a real chance to be one of the best things to happen to regular people, not just the people who already had the most. There will always be people upset no matter what happens, that is true of every major shift in human history. But I think the majority ends up better off, and what will matter most through all of it is staying human to human wherever you can, not losing that thread while everything else changes around it.

This stuff is clever and it is not aware, at least not that anyone can show. It started as something that predicted the next word and it is becoming something that generates original ideas and solves entire disciplines. That is worth taking seriously without taking the doom or the hype at face value. Watch out for yourselves.

This is the same discipline behind why I keep telling business owners to run the Answer Engine Optimization audit before AI starts making decisions about them without their input, and why I tell every operator to push back on AI output before publishing it. Same posture, different room. Assume the machine is involved, then act deliberately instead of by accident.

Set up that safe phrase tonight. Watch out for yourselves. I'm Connor with honor. Be well, and I will see you in the next one.