An AI first and foremost is a Probability Engine based upon weather prediction models

As  programmer myself the easiest ways to finance and develop AI would have been first for the Government to develop better weather prediction models through programming worldwide. It allowed the Big AI companies now and in the future to benefit from these extreme prediction models basically covering All points in the world with their predictions.

However, just like predicting the weather it is not a perfect science and more like Gamblers in Las Vegas betting on different things. So, the best way to understand AI is to think of it as "Prediction Models" or Probability Models, It's not rocket science yet but AI likely would be involved in Rockets too at this point and the design of most or all things of this nature.

Is this good or safe? Not completely because prediction models can often be wrong. They might be 80 to 90% right most of the time.

But, do you really want to drive your car built by a prediction model?

Or fly in a Boeing Plane built by a prediction model.

It could say: "I predict this is going to fly."

That means that 80 to 90% of the time the thing might fly or drive safely!

This isn't something you would want to ride in or fly in if you value your life!