How could Artificial Intelligence become dangerous?

[California]

What I really worry about is how an ever-growing government will recover lost tax revenue from all the AI instances which will replace today's working man.

Simple - tax AI, if that's what has become the means for generating income for corporations.

 

[California]

Here's an article I found saying increased use of AI is already overloading the power grid in some places and it will get worse. The author thinks China has avoided this issue with their emphasis on long term planning.

AI experts return from China stunned: The U.S. grid is so weak, the race may already be over

China is “set up to hit grand slams,” longtime Chinese energy expert David Fishman told Fortune. “The U.S., at best, can get on base.”

fortune.com

 

[Hay Dude]

Here's an article I found saying increased use of AI is already overloading the power grid in some places and it will get worse. The author thinks China has avoided this issue with their emphasis on long term planning.

AI experts return from China stunned: The U.S. grid is so weak, the race may already be over

China is “set up to hit grand slams,” longtime Chinese energy expert David Fishman told Fortune. “The U.S., at best, can get on base.”

fortune.com

Or their long term emphasis on building coal fired plants while guilting us into buying their junk solar panels and windmills.

China laughs at the stupidity, guilt driven madness our country is undergoing.
Hopefully coming to a stop rapidly.

China:
“Coal use isn’t cast as a sign of villainy, as it would be among some circles in the U.S.”

 

[riptides]

Automated AI + drone tech used in Ukraine could result in whoever controls the machines controls the world.

There are serious off switches yet to be introduced in the drone wars.

 

[riptides]

.....

In summary Zuckerberg's Meta offers imaginary friends to the desperately lonesome. While the quality of advice, 'friendship' baked in to this version of AI is harming a lot of people. They're far worse off compared to never heard of it. The 'celebrity girlfriend' scams are tame. compared to this.
....

This is the future, if more people live their life entirely online.

Bladerunner 2049.

I can see interactions in the medical field and elder care.

Not all is bad.

 

[California]

Bladerunner 2049.

I can see interactions in the medical field and elder care.

Not all is bad.

Agree, not all. As a research tool AI is already far ahead of a specialist researching something. For example prior legal decisions that establish precedent for an issue. It has a far larger universe of data available, and it can scan it quickly. But I have read of it making up citations to rulings that never existed. Caveat Emptor. Similar brilliance, and potential issues, in the medical field.

A mundane example: Last week I asked Google the interior cargo capacity with rear seats folded. for three SUV's. Two cars over 20 years old and the third, a current model. The answer came back as soon as I hit enter! No researcher can find stuff that fast. And it illustrates that this obscure forgotten data is now recorded and available at its 'fingertips'.

I think the problem is going to be when people rely on AI-generated conclusions where it doesn't know its source material is unreliable. One example I read about was Elmer's Glue suggested, when it was asked how to make pizza topping less runny. Quoted from a joke post somewhere but it didn't know the difference. And this morning I read of a couple stranded at the airport after they were advised they didn't need visas for a country - but they did, and the airline wouldn't let them board. Now consider the political world in the light of these examples. ....

The next generation is going to inherit a world unlike ours.

 

[Runner]

Reading the local weather forecast discussion this morning, it is loaded with statements like this:

Model guidance has struggled to resolve the
short term forecast.

While low res guidance has been picking up on the multitude of short
waves across the pressure levels, high resolution CAMs (computer aided models) have been
doing their best to try to resolve all of these interactions to
varying degrees of plausibility.

Model guidance is really struggling to resolve the multiple layers
of boundary interactions that are ongoing.

Long term guidance keeps this cooler air mass around through the
rest of the week with the next notable chances for precipitation
coming towards next weekend.

Anytime they use the word "guidance", it means they are looking at computer models. At least they are admitting that computers are doing most of the forecasting now.

 

[Jstpssng]

I view AI the way that I do Wikipedia...
I skip right over it. Especially when it comes to medical questions I want to know the source.

 

[CloverKnollFarms]

Those under 25 go to it first, and trust it more. That is the reality. Us older folks, we may have our view, but in short order the world won't be ours. Our thoughts on this particular topic are quite moot.

 

[Torvy]

Simple - tax AI, if that's what has become the means for generating income for corporations.

It is impossible for a "corporation" to pay taxes. They are just a pass-through for taxes on regular people; a misdirection play for higher taxes which naive people believe are paid by someone else...rather like tariffs.