AI Everywhere From Classrooms To Battlefields

Yesterday, in New York, the heads of the biggest AI companies sat down in front of the United Nations Security Council.

That’s the room where the world’s nations take up war and peace. OpenAI’s chief, Sam Altman, spoke there in person. Anthropic briefed the Council too. When a subject gets called into that room, it has stopped being a technology story. It has become a question of who keeps the peace.

One of the men who spoke was Yoshua Bengio. He’s one of the scientists who built the foundations of modern AI, and he has spent the last few years warning about what came out of his own work. His message to the Council was simple. AI that gets beyond human control is a threat no country can contain alone, and it pays no attention to borders.

The UN’s own summary said something else I want you to hear. It said AI is now being used everywhere from classrooms to battlefields, with few guardrails in place. Think about the span of that. Your grandchild’s homework and a weapon in the field, and in between, very little holding either one to a rule.

Two days before that meeting, the UN’s new scientific panel on AI put out its first report on AI agents. Those are the kind that don’t just answer questions but go out and do things on their own. The panel called for stronger safeguards, and warned that the protections we have now are starting to come apart. So the warning didn’t come from one man. It came from the scientists, and then it came to the table where nations sit.

This same week, the President of the United States and the President of China are sitting down to talk. Bloomberg reported that on AI, the two of them agree on one thing. Neither side feels it can afford to slow down. Washington and Beijing, the report said, seem more worried about who wins the race than about whether anyone can steer the car.

So picture it the way I do. On Wednesday, the world’s peacekeeping table heard the warning in plain words. By the end of the week, the two drivers with the most horsepower on the road agree to keep their feet on the gas. The warning was heard. Then it was filed.

I spent a lot of years on job sites and around heavy equipment. You learn that the most dangerous man on a site isn’t the one who doesn’t know the danger. It’s the one who knows it and figures it’ll be somebody else’s problem. Every accident report I ever saw had a line in it where somebody knew and kept going.

I don’t say that to blame the men at the top for everything. They’re caught in a race, and each one believes that if he slows down, the other one wins. That’s a real fear. But a race where both drivers agree nobody brakes is how a pileup happens. And the people in the pileup are never the ones who set the pace.

That’s you. You’re the one who didn’t get a seat at the Security Council. You didn’t get invited to the summit. But you’re the one whose child is using this in a classroom, and whose job, pension, and power bill are riding on how this goes.

So what can a regular person do when the big table hears the warning and drives on anyway? More than you might think. The safeguards that are coming apart at the top can still be held at the bottom, one room at a time.

Start with this. When you use one of these machines, you set the rules before the work starts. Tell it what it may do and what it may not. Tell it to say so when it doesn’t know. Tell it to show you where it got something, so you can check. A rule you set in your own room is a guardrail nobody in Washington or Beijing can take away from you.

Then keep a record. Write down what the machine told you and what it did. It doesn’t have to be fancy. A notebook works. When something goes wrong, the person with the record is the person who can prove it. The whole world is short on records right now. Don’t be.

And last, talk about it. Not in a panic. Just plain, at the kitchen table, at church, over the fence with your neighbor. The men at the summit read the room like anybody else. When enough ordinary people start asking who’s watching the machine, that question finds its way up to the table.

I’ve written before that an AI should carry a rank, not a crown, like Major Tom answering to Ground Control. This week the world got a good look at what happens when nobody on the ground is holding the line. The warning went up. The answer that came back was full speed ahead.

Ground Control isn’t a room in New York. It’s you. It’s every person who sets a rule, keeps a record, and asks a question before trusting the answer. They heard the warning and kept driving. You don’t have to.

Load this file before each AI chat session.
the-faust-baseline-public-release-1Download

This post was drafted with AI governed assistance and reviewed and directed by Michael S. Faust Sr. before publication.

” Attic Thoughts”-library – Intelligent People Assume Nothing