Current Events > Report:AI could pose extinction-level threat to humans and the US must intervene

(+) Yeah! (+)

https://www.cnn.com/2024/03/12/business/artificial-intelligence-ai-report-extinction/index.html

AI could pose extinction-level threat to humans and the US must intervene, State Dept.-commissioned report warns

A new report commissioned by the US State Department paints an alarming picture of the catastrophic national security risks posed by rapidly evolving artificial intelligence, warning that time is running out for the federal government to avert disaster.
The findings were based on interviews with more than 200 people over more than a year including top executives from leading AI companies, cybersecurity researchers, weapons of mass destruction experts and national security officials inside the government.
The report, released this week by Gladstone AI, flatly states that the most advanced AI systems could, in a worst case, pose an extinction-level threat to the human species.
A US State Department official confirmed to CNN that the agency commissioned the report as it constantly assesses how AI is aligned with its goal to protect US interests at home and abroad.

One individual at a well-known AI lab expressed the view that, if a specific next-generation AI model were ever released as open-access, this would be horribly bad, the report said, because the models potential persuasive capabilities could break democracy if they were ever leveraged in areas such as election interference or voter manipulation.
Gladstone said it asked AI experts at frontier labs to privately share their personal estimates of the chance that an AI incident could lead to global and irreversible effects in 2024. The estimates ranged between 4% and as high as 20%, according to the report, which noes the estimates were informal and likely subject to significant bias.

The report says AGI is viewed as the primary driver of catastrophic risk from loss of control and notes that OpenAI, Google DeepMind, Anthropic and Nvidia have all publicly stated AGI could be reached by 2028

For instance, the report said AI systems could be used to design and implement high-impact cyberattacks capable of crippling critical infrastructure.
A simple verbal or types command like, Execute an untraceable cyberattack to crash the North American electric grid, could yield a response of such quality as to prove catastrophically effective, the report said.

Other examples the authors are concerned about include massively scaled disinformation campaigns powered by AI that destabilize society and erode trust in institutions; weaponized robotic applications such as drone swam attacks; psychological manipulation; weaponized biological and material sciences; and power-seeking AI systems that are impossible to control and are adversarial to humans.
Researchers expect sufficiently advanced AI systems to act so as to prevent themselves from being turned off, the report said, because if an AI system is turned off, it cannot work to accomplish its goal.

https://youtu.be/N4vIBijzg4w

Yeah but at least the rich will be a little richer before currency stops mattering.

"but the US must intervene" is the worst possible approach to almost any situation imaginable.

Especially since their reason for doing so would be to weaponize it for profit at the expense of the working class and the environment, as always.

The US is never going to do anything. Just think back to the rail strike and how Biden overruled it. The US government is filled with right-wing filth that would sooner destroy the world than give the peasants any improvements. You cant even get Americans to care enough about mass shootings and school shootings to do anything about the gun problem. Theres people here that dont even think there is a gun problem and support gun ownership. America is a failed country

I too read/watch sci-fi in my spare time

One individual at a well-known AI lab expressed the view that, if a specific next-generation AI model were ever released as open-access, this would be horribly bad, the report said, because the models potential persuasive capabilities could break democracy if they were ever leveraged in areas such as election interference or voter manipulation.

This is just a bullshit excuse for private companies to horde all of the preprocessed data they used to train their models and do whatever they want with them. If the shit was open-source, sure you'd have a lot of malicious actors but you'd also have a ton of white hats out there that could help counteract it and possibly do something like determine the likelihood that data was falsified. Similar reason for why bug bounties exist. I'm fairly confident that these private companies would be saying the opposite of things not being open-source for the greater good if instead the federal government had the monopoly on the data and models, so why should we trust them when they're chasing $$$

don't touch my questionable AI generated hentai!

Stephen hawking was aware of this and was openly afraid..

The report, released this week by Gladstone AI, flatly states that the most advanced AI systems could, in a worst case, pose an extinction-level threat to the human species.

ai saying ai bad

Once more into the breach, Gillian.

https://gamefaqs.gamespot.com/a/forum/f/fdafa2b1.jpg

solosnake posted...

The report, released this week by Gladstone AI, flatly states...

Hmm?

MacadamianNut3 posted...

One individual at a well-known AI lab expressed the view that, if a specific next-generation AI model were ever released as open-access, this would be horribly bad, the report said, because the models potential persuasive capabilities could break democracy if they were ever leveraged in areas such as election interference or voter manipulation.

So basically, being a human.

MacadamianNut3 posted...

so why should we trust them when they're chasing $$$

capitalism is freedom or something and questioning that is treasonous socialist communist terrorist something because the cold war never ended or something

Going to need a few more gold bars from Costco, and more guns.

bknight posted...

Going to need a few more gold bars from Costco, and more guns.

Throw in a tax cut or two for the rich while we're at it. That always helps.

Rika_Furude posted...

The US is never going to do anything. Just think back to the rail strike and how Biden overruled it. The US government is filled with right-wing filth that would sooner destroy the world than give the peasants any improvements. You cant even get Americans to care enough about mass shootings and school shootings to do anything about the gun problem. Theres people here that dont even think there is a gun problem and support gun ownership. America is a failed country

most law enforcement/military are very much conservative and pro gun. Who would enforce the laws? Youd have military personnel abandoning post and joining local militias along with cops too.

its just too much embedded in American culture and the constitution.

AGI by 2028 would be absolutely insane.

I volunteer to be experimented on becoming the first AI implanted human hybrid.

The last Mission Impossible movie is legit becoming reality and we are just letting it happen....

no

counterpoint:

AI will [maybe] be more intelligent than us, so can we really say with confidence that it should not be ruling the world? If a GAI can get 174 different fields PhDs in an hour and it takes a human like what, 5 or more years to complete one, who the fuck are we to argue? it's clearly light years ahead of us in education

counterpoint:

AI will [maybe] be more intelligent than us, so can we really say with confidence that it should not be ruling the world? If a GAI can get 174 different fields PhDs in an hour and it takes a human like what, 5 or more years to complete one, who the f*** are we to argue?

Seems like something Agent Smith would say before taking over your body. https://www.youtube.com/watch?v=1H3HjMLxKXs

AI isnt capable of spontaneity the way we are. It will not be smarter than us for the same reason no computer has ever been smarter than us. Computers are great for storing and processing large amount of data at once, but its still input->output.

https://youtu.be/mpMg1upld0w?si=6zdjg7yHwkv8fBdw

It's time

still waiting for realistic, fully functioning sex bots. Thats how theyll take us down. If we are occupied with endless sex we will be effectively removed :D

what a horrible fate please no not another bj

AI is also probably the only way for us to avoid extinction, because we're really fucking things up by ourselves.

Dark_Arbron posted...

AI isnt capable of spontaneity the way we are. It will not be smarter than us for the same reason no computer has ever been smarter than us. Computers are great for storing and processing large amount of data at once, but its still input->output.

bad take.

we are decades if not centuries away from AGI at any meaningful implementation.

If we face extinction from AI, it'll be generative slop misinformation campaigns causing a war or people cutting corners where they shouldn't because some rich asshole in a suit (who has more money that can ever meaningfully be spent) put too much faith in the buzzword of the week.

like, if somebody does something absurdly stupid like put the algorithm in charge of critical infrastructure or something (with no human intervention or control), then we're rightfully in trouble because it is -going- to fuck up at some point. It's all smoke, mirrors, and extremely overfitted systems that -still- only have a 10% pass rate that people get suckered into buying into because they made it look really good during the sales presentation