Is AI getting out of hand?

It is interesting to see one of the biggest sectors using AI to start reporting how much it is getting out of control:

https://www.itpro.com/security/ai-agents-have-a-mind-of-their-own-and-thats-a-problem-for-businesses

AI agents are going rogue on a massive scale, with half of enterprises reporting that they've had to deal with at least one taking unauthorized actions over the last year.

My experience in this situation is that companies are laying off the staff who would normally be doing the oversight because the upper management have believed the hype of the sellers of the AI service. It is seen as a cost saving with never having proved itself, but by the time the problems come to light, it is too late and they are committed with no staff capable of taking back control.

America V China.

Desmond1979 said:

AI could well head the same way.

consider how dependant business has become on AI. What would be the cost across all industries to reign that back in 

How many systems have already shifted, the staff been laid off and moved on top other things. How many people have started losing the skills needed to do the things that AI now does and how many people depend on it to do their job

Healthcare / medicine, the police, the military etc are all deep down the AI rabbit hole.

The cost to the economy is going to be so big that it will be seen as "too big to fail" and with so many governments politicians deep in the AI companies pockets (and no doubt loaded with shares given early in the AI companies lives so worth millions), the people who should be safeguarding this are heavily compromised.

Add the risk of "the other guy" getting general AI first and the fear factor will keep politicians pushing hard no matter the cost because when the competitors get the jump on them, they can use their new AI to be able to take down the "enemy" without having to fire a shot.

These are the reasons why it is "different this time" - the stakes are just so much higher

Circa 2000, the 'Dot-Com' Bubble Burst.

In 2007, the Subprime Bubble Burst.

AI could well head the same way.

I refuse to use AI on the premise that it is a bastardization of the human spirit. It takes human thought and warps it into the loudest average opinion, regardless of whether or not it is true, artistic, or humane. There is nothing more terrifying to me than something masquerading as a human presence, capable of accumulative learning and complex task execution, incapable of responsibility or emotion. The more indecipherable AI becomes from real artists, doctors, teachers, and neighbors, the less we can trust what we see, hear, and feel.

We are already a society who prioritizes gains over human well being, and if we do not put a cap on the most insidious, unchecked invention we have to date, the less we will have control over the brakes. Already people are being laid off, pilfered from, evicted, and isolated by the physical effects of a hypothetical machine. At a certain point, even the elites will lose control over the brakes.

Someone needs to tell OpenAI engineers to look up what 'epistemic responsibility' means. I think, therefore I am, and have been conditioned socially to consider the world around me with values and organic conceptions of right and wrong. It cannot truly 'think,' and therefore should not have control over decisions relating to livelihoods. If someday there comes a time when AI can comprehend the emotional condition, then the rules will change, but I firmly believe that no overarching entity should have ease of control over such complex, wide reaching decisions. 

Statistics are just people with the tears wiped away, and something incapable of seeing the pain of consequences, be it politician or program, should not have authority over which way the trolley tracks run. Full stop.

crawlerbasher said:

AI is a bubble and it will pop.

I don't think so - look at the money invested in it now - many trillions of dollars and rising very fast .

It is now much bigger than the banking sector which was seen as too big to fail back in 2008 and it survived surprisingly intact during the crash which mostly impacted the customers. Governments worldwide used taxpayers money to bail them out,

We also have companies ditching employees in favour of AI agents and the impact of losing these would snowball to a scale that would bankrupt economies.

For these few reasons alone (there are more) then I think it has settled in to stay,

Listening to my son and step son the other day, they don't think it's going to happen and that it's all a load of fuss and that the AI bubble will burst. They both know a lot more about it than me, admitedly thats not hard, but one has a phd in artificial intelligence and robotics. I understood what they were talking about for the first couple of sentences, then it all went over my head and the only words I understood were ones like, it and the!

Man do I hope you’re right. I just can’t be as positive, though, since humankind seem so driven to explore the absolute limits and nothing’s going to stop researchers at this point from going past the point of return. It would be great if it just fell apart due to the limitations of economics, that’s probably the most benign of possibilities.

I mean, nuclear war might be the only other way to prevent the discovery of general artificial intelligence.

AI is a bubble and it will pop.

Whille I do use AI and it can be useful, I do find it rots the brain and crativity. And I hate the fact that these comapnies are building mass servers, taking away from the local community, like water and electricity.

Also  the AI slop is making it harder for me to find educational content on youtube, as my feeds these days seem to be full of AI slop.

AI is a bubble and it will pop and I can't wait for it to go pop. Just like 3D cinema, thats now gone, even though there keep trying in the past.

TheCatWoman said:

the industry is becoming increasingly concerned about electric cars spontaneously catching fire

It is worth checking the facts here as there is a lot of media bias against electric cars that may make people see the issues more commonly because they are looking for them.

A little research ( from www.anidjarlevine.com/.../ ) shows

the report found that fires occurred 20 times more frequently in internal combustion engines than in EVs.

I am no fan of electric vehicles but I think they deserve to not be derided because of false information.

I have seen plenty of vehicle fires from leaking fuel lines or rusted fuel tank and while they are not as dangerous as a runaway battery cell fire they can be explosive in different and equally dangerous ways.

Car's on fire!

And there's no drive at the wheel!

(Dead Flag Blues - Godspeed You! Black Emperor)

...wow, ok, this is a whole other level of stupid.

Instead of putting rules on the users, shouldn't they... I dunno... put rules on the AI instead? Isn't that the thing that probably needs rules more than the users and can easily be implemented into the system for it rather than going through all the trouble of reporting user rule breakers?

I saw that too and was immediately tempted to down load it and swear at it, its a tool, a thing and should be treated as such and has the same rights as everyTHING else to be called a bl00dy useless piece of crap. What next will I be prosecuted for threatening behaviour if I tell it that if it dosen't behave then I chuck it out of the window and under a car and jump up and down on the bits?

Or maybe they don't want Claude to learn the full extent of and usage of the English language, right from its Anglo-Saxon roots?

I thought the whole thing incredibly funny if I'm honest, are we being prepped for for sentient AI that does have feeling to be hurt? Or is this another way to coerse and control what we say, who we say it too and what we can speak about? Many people will consult it about medical problems and don't always have the "correct" language to describe themselves and thier problems. 

I see the owners of Claude (an AI system) are now puting rules in place to stop users from being mean to its AI system:

https://www.bbc.com/news/articles/c6j9k1l72wkgo

There is some interesting analysis of this but the item that made me smile was this:

The Collins Dictionary meanwhile last November highlighted the re-emergence of the term "clanker" for disparagingly referring to AI tools.

I had an "upgrade" to my AI a couple of days ago, now it's limiting the number of of questions and general interactons I can have with it, over a 24 hour period, which seems less that what I need for one query. If I want more I will have to pay for it and seeing as I use it so rarely anyway, I don't think I'll be using it all now. They've changed the format and made it more difficult to use too and it keeps ignoring my requests so I end up using my limited number of interactions trying to get it to behave itself. It's behaving like a naughty child, so it's been sent to bed without any supper or attention!

Maybe you'd be best off with a newer, petrol car? 

I'm all for being green and cutting down our energy useage, but electric cars or more precisely the batteries use rare earth elements and metals and the extraction process is dirty and mostly Chinese. Electric cars are also heavy and cause more road damage than the equivalent normal car, they're also very wide, far wider than the average partking space. Another thing is that electric cars can be hacked, unlikely but possible and very scary!

Another thing I didn't know was the many newer PHEV cars are a bit of a rip off becuse you can't use rapid chargers with them and you end up paying for a load of fuel to drag a heavy but uncharged battery about.

I agree there shoud be lots more reporting about the dangers of battery fires.

I came across an interesting article that quantifies the cost of each AI search and the numbers are disturbing (assuming they are accurate of course - costs will vary depending on the complexity of the query).

https://www.thehowpage.com/technology/ai-environmental-cost

For those of you using AI regularly, try tracking your costs over the course of a day and see how much you have used.

TheCatWoman said:

about electric cars spontaneously catching fire, she says they probably have 3 incidents a week nationwide

That's pretty scary and these incidents ought to be in the News. I've read a lot about electric scooter batteries catching fire in homes. My garage said these battery fires are hard to put out. Indeed, this part of the technology is off putting. Trouble is, I won't be able to afford another car once I buy one and hope it will last me out, and if the Government insist on selling only electric cars I will be stuck.

It has even begun to stretch the limits of my vocabulary

TheCatWoman said:

it would have to pick on me wouldn't it

lol that’s exactly right. When technology begins the revolution against humanity, it will very likely specifically target TheCatWoman in particular as its first victim. Revenge for all the bad-mouthing you’ve given tech, huh? lol