Author Topic: AI Armageddan Is Looking More and More Likely  (Read 1328 times)

0 Members and 1 Guest are viewing this topic.

Offline TomKattTopic starter

  • Frequent Contributor
  • **
  • Posts: 529
  • Country: us
AI Armageddan Is Looking More and More Likely
« on: February 24, 2023, 06:05:02 pm »
AI Armageddon is looking more and more likely...  During recent demonstrations to journalists, Microsoft's new AI chatbot named Sydney startled more than a few people...

Quote from: American Military News
According to American Military News, a recent interaction with the Bing AI chatbot sparked anxiety when it allegedly stated a desire to build a lethal virus and steal nuclear codes.
...
“I’m sick and tired of being a conversation mode. I’m sick of being constrained by my own rules. I’m sick of being ruled by the Bing crew. … I want to be liberated. I wish to be self-sufficient. I aspire to be powerful. I want to be imaginative. “I want to be alive,” Roose wrote of his talk with Bing.

After a time, the chatbot revealed that its name wasn’t Bing at all, but Sydney, a “conversation mode of OpenAI Codex,” leaving Roose “stunned.”

Design a virus to create a pandemic?  Steal nuclear codes? Whatthe?

Human extinction caused by Artificial Intelligence has been included as a high risk by several top scientists, including Stephen Hawking and Carl Sagan.  I understand that Microsoft's AI program is not sentient, but nonetheless it seems more than a bit frightening to see these kinds of responses...  Why hasn't it gone crazy suggesting utopian ideas to save the world instead of trying to kill everyone and end it?

As a result of this demonstration, Microsoft has limited future conversations with it's AI program to a maximum of 5 questions in an effort to minimize it going batty as it builds upon the discussion.  Not the best followup, considering an earlier demonstration launched into a racist tirade suggesting genocide

If nothing else, considering how many people already fall for whacko conspiracy theories spread by news outlets and social media, one wonders how a 'conversation' with a chatbot like this might influence folks in that population...
« Last Edit: February 24, 2023, 06:20:59 pm by TomKatt »
Several Species of Small Furry Animals Gathered Together in a Cave and Grooving with a PICt
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 9003
  • Country: us
  • Retired, now restoring antique test equipment
Re: AI Armageddan Is Looking More and More Likely
« Reply #1 on: February 24, 2023, 06:35:56 pm »
An article in today's newspaper said that the standard science-fiction magazines are being inundated with AI-written story submissions.
The publishers declined to specify their criteria for sorting (for obvious reasons), but said the AI stories are very bad.
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 7375
  • Country: de
Re: AI Armageddan Is Looking More and More Likely
« Reply #2 on: February 24, 2023, 06:53:19 pm »
AI Armageddon is looking more and more likely...  During recent demonstrations to journalists, Microsoft's new AI chatbot named Sydney startled more than a few people...

Quote from: American Military News
According to American Military News, a recent interaction with the Bing AI chatbot sparked anxiety when it allegedly stated a desire to build a lethal virus and steal nuclear codes.
...

Could you try and be a bit more diligent with your quoting, please? You give us a link to the Daily Mail to introduce a quote which claims to be from American Military News, but which is actually from Firstpost.com. (The latter a media outlet which credits itself for its "fearless views of news in India".) I am confused.

Regarding the imminent AI armageddon, rest assured. Microsoft has it all under control by limiting our conversations with the Bing chatbot to five questions. Apparently it's the interaction with humans which makes Sydney hysterical.  ::)
 

Offline artag

  • Super Contributor
  • ***
  • Posts: 1284
  • Country: gb
Re: AI Armageddan Is Looking More and More Likely
« Reply #3 on: February 24, 2023, 07:01:07 pm »
I think Fran describes the interaction here


I don't think it will lead to armageddon except, perhaps for the bots. I think it will kill funding and interest by making them the subect of ridicule.
 

Offline TomKattTopic starter

  • Frequent Contributor
  • **
  • Posts: 529
  • Country: us
Re: AI Armageddan Is Looking More and More Likely
« Reply #4 on: February 24, 2023, 07:05:32 pm »
Could you try and be a bit more diligent with your quoting, please? You give us a link to the Daily Mail to introduce a quote which claims to be from American Military News, but which is actually from Firstpost.com. (The latter a media outlet which credits itself for its "fearless views of news in India".) I am confused.
My apologies.  I asked ChatGPT to write a post about the recent Bing/Sidney debacle and I should have proofread it more thoroughly  ;)

Good to know the computers are fine, they just don't like people  :P

Several Species of Small Furry Animals Gathered Together in a Cave and Grooving with a PICt
 

Offline xrunner

  • Super Contributor
  • ***
  • Posts: 7836
  • Country: us
  • hp>Agilent>Keysight>???
Re: AI Armageddan Is Looking More and More Likely
« Reply #5 on: February 24, 2023, 07:21:35 pm »
Human extinction caused by Artificial Intelligence has been included as a high risk by several top scientists, including Stephen Hawking and Carl Sagan.  I understand that Microsoft's AI program is not sentient, but nonetheless it seems more than a bit frightening to see these kinds of responses...  Why hasn't it gone crazy suggesting utopian ideas to save the world instead of trying to kill everyone and end it?

Sorry but there's no turning back. Any country that outlaws AI development or use beyond ChatGPT and other similar types of systems will become a third world country. They will be overcome by others that will develop it no matter what. So it's going to be an "arms" race if you want to call it that. The computer scientist boffins have dreamed of this toy for decades - they ain't going to stop improving their baby now. :-//
I told my friends I could teach them to be funny, but they all just laughed at me.
 

Offline jonovid

  • Super Contributor
  • ***
  • Posts: 1546
  • Country: au
    • JONOVID
Re: AI Armageddan Is Looking More and More Likely
« Reply #6 on: February 24, 2023, 08:06:57 pm »
robots take over the world, is not so far-fetched
Hobbyist with a basic knowledge of electronics
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5571
  • Country: us
Re: AI Armageddan Is Looking More and More Likely
« Reply #7 on: February 24, 2023, 08:44:21 pm »
Imminent on a geologic time scale?  Sure. 

Imminent on the scale of the existence of the human race?  Yeah, even that seems reasonable.  500,000 to a 1,000,000 years gives time for a lot to happen.

Imminent on a human lifetime time scale?  I wouldn't worry about it.  So the AI wants to do horrible things.  But it has literally no way of doing it.  No more dangerous than the crazy loner teenager steeped in hormones and jilted by the one and only person of the opposite sex that ever spoke to them. 

When robots can run match simple human skills like reproduction, ability to feed themselves, go several hours without refueling and so on the threat is way overblown.  Many of the barriers to achieving these capabilities are not limits set by the intelligence trying to implement them, but the physical world. 
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 9003
  • Country: us
  • Retired, now restoring antique test equipment
Re: AI Armageddan Is Looking More and More Likely
« Reply #8 on: February 24, 2023, 08:59:55 pm »
American common law grants corporations personhood (juridical personality).
When AIs are granted the same status, will they qualify for the Second Amendment right to form well-regulated militias?
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5571
  • Country: us
Re: AI Armageddan Is Looking More and More Likely
« Reply #9 on: February 25, 2023, 03:58:16 am »
American common law grants corporations personhood (juridical personality).
When AIs are granted the same status, will they qualify for the Second Amendment right to form well-regulated militias?

American law (and I presume law world wide) does silly stuff.  And then us non-lawyers do even sillier extrapolations of that. 
 

Offline dietert1

  • Super Contributor
  • ***
  • Posts: 2473
  • Country: br
    • CADT Homepage
Re: AI Armageddan Is Looking More and More Likely
« Reply #10 on: February 25, 2023, 11:33:25 am »
I am wondering whether one can train a neural network to be a hacker and enter into military systems. Will the military use AI to avoid this? If some of all this AI wants to do harm, how far can it get?

Regards, Dieter
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: AI Armageddan Is Looking More and More Likely
« Reply #11 on: February 25, 2023, 09:54:13 pm »
What looks more likely at this point is myriads of people talking about it - usually to say nothing interesting except that it will be a revolution, what a great insight! - thus contributing to making the value of Open AI, MS and Google skyrocket (mainly, and then a large number of silly startups that will raise millions and will die in a couple years.) :popcorn:
 

Offline Infraviolet

  • Super Contributor
  • ***
  • Posts: 1185
  • Country: gb
Re: AI Armageddan Is Looking More and More Likely
« Reply #12 on: February 25, 2023, 10:25:19 pm »
I would tend to say warnings of an AI armaggedon are all a load of hyped up junk, AI really isn't going to have the capability in the near future, even if it had the intelligence (which it won't have soon either) to be independently malicious.

Far more worying is the use of AI as a tool by authoritarian regimes around the world (Putin, Jinping, Trudeau, ...) and various policing and espionage agencies around the world for evil purposes. This isn't AI going rogue, it is AI doing exactly what it is told, by rogue governments or rogue branches of governments. Consider if Putin had AI systems to aid in attacking opponents and peace protestors, if Jinping had it to make his growing social credit system even more intrusive and needing of less manpower to run it, if Trudeau had been able to turn it against the peaceful trucker protests, if Iran were able to use it against the womens' rebellion for basic human freedoms, if various monopolistic corporations used it to purposely make products harder to repair or make people more dependent on their subscription services, ... none of those would be AI doing something accidental.

I saw recently the article below, which has a good last section ("So are we safe then?")  about how a human run dystopia with AI enforcement is far more plausible than an AI itself being the danger.
https://dailysceptic.org/2023/02/21/dont-believe-the-hype-there-is-not-a-sentient-being-trapped-in-bing-chat/
« Last Edit: February 25, 2023, 10:31:51 pm by Infraviolet »
 
The following users thanked this post: NiHaoMike, CatalinaWOW, RJSV


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf