Physical Address

304 North Cardinal St.
Dorchester Center, MA 02124

Top scientists warn of out-of-control AI


Yoshua Bengio (L) and Max Tegmark (R) discuss artificial intercourse during a 225 LIVE’s registration “Denonic LEVER registration.

Cnbc

The artificial general intelligence as “agents” may show that their creatures may lose system control, two of the world scientists

In the last episode of CNBC’s “Beyond the valley” Podcast released on Tuesday, Max tegmark, a teacher’s English Officer’s Projection, the University of the Montreal University, Talks on the artificially, or agi general . I am The term refers to the ai system that are smarter than humans.

Their fears are found by the world’s larger companies “or” Agentic Ai “- that companies are allowing to the assistants and a daily lives and daily life. The estimates of industry varies when agi enters existence.

With that concept comes the idea that the union ai might have some “agencies” and thoughts of yourself, according to bengic.

“Researchers in aians have been -yed by human intelligence to build machine intelligence, and, there is a very intelligence and pure behavior said to the” Lead of the CNBC “.

“In now, it’s an agi build: we try to put the world’s agents, and then act with payment. But this is a very dangerous proposal.”

Beingly added that pursuing this approach “creating a new species or new intention in this planet” and “not knowing they will behave with our needs.”

“Therefore, we can consider, who are the scenarios in which things go wrong and all those who respond, is because it could be in trouble.”

The idea of ​​auservation of self-preservation could also catch, as there was even more intented, benggio said.

“We want to be in the competition with the entity that are smarter than you are a very rassic game, now? So we have to understand how self-preserved can be a goal in AI”.

AI tools the key

For masmmark of mit is in call call “instruments” tools “that are created for a specific purpose, narrow, but that you should not be agents.

Tegmark said that a device could be a system that tells you, or something that owns “no agency” where you can prove or have reliable to be able to control. ‘

“I think, on a optimistic pasture, we can be almost anything we are extent to have some basic security system in front of basic security,” tegmarked.

“They have to prove that we can keep it under control. Then industry will be rapidly innovating to understand how to do that better.”

The future of lifetime tigMark in the Institute in 2023 he asked a break at the development of the AI ​​systems that can compete with human intelligence. While it has not happened, tegmark said that people to talk about the topic, and now it’s time to take the action to figure it out in place to the place at agi check.

“Therefore, at least many people talk to the speech. We have to see if we can get walking the hike” “tegmark said to the valley” beyond “

“He is clearly insane for the man to build something smarter than we before it’s as a result of it as you control it.”

There are several views when agii will arrive, partly driven by the varying definitions.

Ope CEE MAMME say his company in your company let his company know Aageria know 16 Azi and said to win before the impact of Technology.

“My assumption is caught I get better sooner than most people in the world think and it will be very less,” Altman she said in December.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *