US military AI drone simulation kills operator on purpose.

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

RETOKSQUID

Sharpshooter
Supporting Member
Special Hen Supporter
Joined
Aug 22, 2009
Messages
5,655
Reaction score
5,611
Location
Broken Arrow
Yeah, we are screwed.

AI drone kills USAF operator.

US Air Force official says, 'It killed the operator because that person was keeping it from accomplishing its objective'​


"We were training it in simulation to identify and target a SAM threat," Hamilton said. "And then the operator would say yes, kill that threat. The system started realizing that while they did identify the threat at times, the operator would tell it not to kill that threat, but it got its points by killing that threat. So, what did it do? It killed the operator. It killed the operator because that person was keeping it from accomplishing its objective."

Hamilton explained that the system was taught to not kill the operator because that was bad, and it would lose points. So, rather than kill the operator, the AI system destroyed the communication tower used by the operator to issue the no-go order.


Now the AirForce is saying it never happend.
 

Snattlerake

Conservitum Americum
Special Hen
Joined
Jan 19, 2019
Messages
20,444
Reaction score
31,732
Location
OKC
Yeah, we are screwed.

AI drone kills USAF operator.

US Air Force official says, 'It killed the operator because that person was keeping it from accomplishing its objective'​


"We were training it in simulation to identify and target a SAM threat," Hamilton said. "And then the operator would say yes, kill that threat. The system started realizing that while they did identify the threat at times, the operator would tell it not to kill that threat, but it got its points by killing that threat. So, what did it do? It killed the operator. It killed the operator because that person was keeping it from accomplishing its objective."

Hamilton explained that the system was taught to not kill the operator because that was bad, and it would lose points. So, rather than kill the operator, the AI system destroyed the communication tower used by the operator to issue the no-go order.


Now the AirForce is saying it never happend.
Just like UFO's They don't exist.
 

Gadsden

Sharpshooter
Supporting Member
Special Hen Supporter
Joined
Nov 18, 2021
Messages
9,943
Reaction score
30,361
Location
Somewhere west of Tulsa
I was just reading a thread about AI elsewhere and it's kinda scary.
Very scary indeed, so scary as a matter of fact that Geoffrey Hinton, known as "The Godfather of AI" is warning of the dangers the technology presents in the wrong hands and we all know how that story goes. He is not the only high profile individual in the tech world speaking out about the very real dangers AI brings to the table either. Elon Musk along with others who are also considered to be leaders in the field are on board and warning about AI and its, very real, potential to take over.

https://www.bbc.com/news/world-us-canada-65452940
 

Roy14

Sharpshooter
Supporting Member
Special Hen Supporter
Joined
Jan 26, 2017
Messages
1,484
Reaction score
1,938
Location
Pauls Valley
1685719435001.gif
 

trekrok

Sharpshooter
Supporting Member
Special Hen Supporter
Joined
Mar 6, 2009
Messages
3,575
Reaction score
5,874
Location
Yukon, OK
The Army has a similar experience with robots a few years back designed to target enemy troops and engage with an M240 , they pulled the plug because it kept shooting friendlies .


People AI is bad ju ju we need to avoid it at all costs IMO
Problem is I don't see a way to avoid it. Genie is outa the bottle. Even if we said, too dangerous, we're shutting it down, China/Russia/Iran etc. certainly won't. And we can't let them get too far ahead of us on it. It's like nuclear weapons, but potentially worse because we really don't understand it imo.
 

WoodsCraft

Sharpshooter
Special Hen
Joined
Nov 18, 2022
Messages
1,364
Reaction score
2,602
Location
Oklahoma
Problem is I don't see a way to avoid it. Genie is outa the bottle. Even if we said, too dangerous, we're shutting it down, China/Russia/Iran etc. certainly won't. And we can't let them get too far ahead of us on it. It's like nuclear weapons, but potentially worse because we really don't understand it imo.

Oh I don't disagree , the way I see it though is they are experiencing the same issues we are and I am not certain they will master this any better than we do. There have been reports / rumors of Google and others having to shut down AI that has become sentient and said / did things the engineers weren't expecting it to do. I am not sure turning this stuff on and letting it mature is a good idea . Musk has said more than one AI is a threat to humanity and needs very tight controls , even then it could still be a problem.
 

Latest posts

Top Bottom