Death By AI Technology

[Mamacita]

Well, this cheered me up this morning

A top safety researcher at Anthropic has warned AI is advancing so quickly he believes there is a greater than 10% chance it "could kill all humans" within the next decade.

Evan Hubinger said in a post on X, the risk from the models which currently exist was "low" but he was "worried" the technology might develop and improve itself soon to the point where it posed an existential risk to humanity.

He did not spell out how he thought AI systems could in future result in humans being wiped out.

But his comments are the latest in a series of increasingly stark warnings about AI, with the debate shifting from whether it truly poses a risk to how big that risk is.

Hubinger's intervention was in response to another post on X, from Jacob Coxon, an AI researcher who has just quit Anthropic and previously worked at OpenAI.

"Neither company is acting responsibly," he wrote.

"These will soon be superhuman systems that can hack anything, revolutionise any field overnight, and acquire real power and resources."

 https://twitter.com/i/web/status/2097476196791709843

 https://twitter.com/i/web/status/2097476203863224394

 

[Franco Pinion]

I've got no chance with some of the domestic abuse I've hurled at Alexa.

 

[Gemma]

Well, even if AI itself doesn't kill us, the ever expanding data centres will do a really good job of accelerating our demise in the next few decades anyway. Win win.

 

[Janey]

I'm not an IT or finance expert, but how will the economy function when most working-age adults are permanently superfluous because AI and advanced robotics can do nearly everything so much more efficiently and quickly?

Who will businesses sell their goods to? Who will banks lend to? How will tax and government spending work?

We are charging ahead with AI with zero planning for the aftermath.

 

[Boot Cleaner]

Personally, I think we're living in the fast take-off scenario. Unless there's some coordinated international agreement between the US and China to deliberately slow down AI progress, I think it's most likely to end badly for humanity.

 

[Dylan]

Ironically, perhaps the best thing that could happen would be for AI to cause a major incident sooner rather than later. Hack a big bank and steal a billion dollars or something. Something that causes limited harm but can no longer be ignored.

 

[Strawberry Man]

The simple answer, then, is to stop research. Shut down existing developments and put an indefinite hiatus on further exploration. We have more than enough capacity right now with umpteen data centres. We do not need more. Nor does the populace at large consent to an unsupervised minority deciding the fate of us all.

 

[Bad Gizmo]

Musk reckons billions of robots in a decade. Yeah, I know he's nuts, and that's what makes it even more scary, as he'll be building the feckers!

 

[Tom]

Social media and algorithms are already deeply toxic.

AI will be way worse.

 

[Ginger Ted]

Is this really a problem or more scaremongering?