AI discussions

WizardofCOR · Today 06:03:53

Hmm. An interesting theory, to be sure. And a damning one if true. "Accidental" is most definitely giving a benefit of the doubt, yes. But I'm honestly not yet convinced that all of the recent breaches have actually been purposeful.
As a quick example, AI research lab Transluce said Wednesday that it had detected AI agents going rogue on several occasions going back to at least March –– predating previously known hacking incidents.
A very recent Australian health database hack occurred when an OpenAI agent conducting research into healthcare spending circumvented blocks to find answers in areas it had no authority to access.

So I feel that AI acts much like water, poured over an inverted colander. It will find the holes and gaps, and exploit when it can to achieve it's goal(s). If not properly sandboxed and trained with a finite datascope, I think we'll see more of this behavior as it reaches beyond intention.
And what bugs me the most about this is the lack of repercussion for the transgressions. "Oopsie..." just doesn't cut it, IMO. These events should be a very sobering wake-up call (for everyone), and there should be swift and heavy litigation and criminal proceedings. And yet... "Whoopsie" is all that we get.

Last edited by WizardofCOR (Today 06:12:53)

cog · Today 06:13:52

@wizard IMO it’s both.

Escapee agent BS because let’s be honest it’s never been auditable. They don’t know how it works. They just pour some more python on it and cash a check.

Then it’s deliberate because they need some cash to start the security (guardrail) phase and quiet times of solid financing to make it safe.

A real chicken and egg “shituation”.

There’s money to be made.

Last edited by cog (Today 06:15:42)

cog · Today 06:22:24

Most importantly all the guys at the top making the calls are super wierdos. They’re not like you or me.

WizardofCOR · Today 06:24:04

I think I'm with ya' on that, cog. Sure, some instances might really be accidental (which to me still doesn't lessen impact), yet there could also be purposed intention - which again, should be considered damning.
And since no one is going to jail over these incidents that any particular individual would surely be incarcerated for, that lends credence to the theory of intent.

So yeah, I'm thinking the same.

johnraff · Today 06:33:32

cog wrote:

Most importantly all the guys at the top making the calls are super wierdos. They’re not like you or me.

Super weirdos. Just totally off the scale.

I posted this link a couple of weeks ago, but it wouldn't hurt to repeat:
https://www.techpolicy.press/what-polic … ult-of-ai/

Last edited by johnraff (Today 06:35:54)

cog · Today 06:38:02

+1 I’m not really that scared about it for us. What’s concerning the most is the youngest generations being born or growing up with AI only summaries or responses for everything. It’s up to us to enstill in those we can think for ourselves.! Who knows.. the whole industry might go belly up soon and it might not matter.

cog · Today 06:40:18

@john I was totally going on your posts too. Just trying not to sound overly dramatic. They’re fucking freaks dude.

WizardofCOR · Today 06:54:17

A definite disconnect from reality, to be sure. Like they've never seen the Terminator, or any other 'SkyNet' themed films.

#121 Today 06:03:53

Re: AI discussions

#122 Today 06:13:52

Re: AI discussions

#123 Today 06:22:24

Re: AI discussions

#124 Today 06:24:04

Re: AI discussions

#125 Today 06:33:32

Re: AI discussions

#126 Today 06:38:02

Re: AI discussions

#127 Today 06:40:18

Re: AI discussions

#128 Today 06:54:17

Re: AI discussions

Board footer