Another Killer A.I. N.e.w.s. Conspiracy

So, I went to Anomalien for a quick scroll when I see this headline:
"Evil A.I. invented 40,000 lethal chemical weapons in six hours!"

(Sighs. Drinks a bitter cup. Copy/Paste post continues.)

Days ago, I read just the opening paragraph of another article concerning an a.i. chatbot user who got it to display a poem featuring their grandmother instructing them on how to make napalm at home. Fun times with first world problems, amirite? Wrong. This is not fun, really.

There are other conspiracy theories on how many times entire continents and communities were destroyed via the magic of careless mortals and their advanced technology. Every day is laden with rumors of doom. Is that the psy-op? Because there is a high percentage of the world population that will declare as fact that human life will continue no matter what factor one can imagine.

I can attest to being raised in the predictive programming of pop-culture, and being a conspiracy theorist because of it.
Seriously, name a few Dystopian fiction products that are a clear and present threat right now.
Name the movies.

I suppose using artificial intelligence in lieu of a compassionate consciousness will murder many as usual. Allegedly.
I've experienced much of what is dismissed as paranormal and or paranoia to worry about death, slow or instant.
What does concern me is everyone else. Just [ATS] members alone are a good sign of "survive and thrive".

These technocratic threats are important, to be sure.
But what is essential is self-control when faced with yet another "disaster conspiracy".
Godspeed to the warriors.

Where do you see yourself in five years?
:Saint:

a reply to: BeNotAfraid

Where do you see yourself in five years?

As part of the collective consciousness.

a reply to: BeNotAfraid

This AI is a generative model called Megasyn not an "evil" AI. It uses different chemical properties in order to predict molecular design using machine learning methods, which are statistical algorithms. It does not tell how to produce the design, which means, scientists would need to figure it out. This AI was not designed to create chemical weapon molecular designs, although it can be used for that. The data scientists simply tested, whether it is possible, by leaving it running it throughout on night.

MegaSyn: Integrating Generative Molecular Design, Automated Analog Designer, and Synthetic Viability Prediction

a reply to: BeNotAfraid

I think a serious overlooked threat is accelerated systemic error rampancy across industries and their subsegments. The indifference or lack of emotion can be programmed around to some degree, but the errors will compound from the bottom up and feed into systemic errors that can't be automated out by other AI. As more AI output is feeding other AI without proper human oversight these minor errors can combine to create rogue waves, popping up in otherwise steady seas (low frequency and amplitude of expected errors). It will rapidly outstrip any kind of manual efforts to keep up with QA/QC, even if somebody was planning to set up a standardized method for measuring it. It's just going to add an element of odd pseudo-randomness to the otherwise human incompetence/error feedback loop. People have been slowly accepting these compromises in quality as human incompetence has been normalized, bugs turned into features, and accountability rendered an outdated concept.

Until AI can understand things and has quantum processing I'm not looking for a John Connor to save us. I think the biggest threat is that AI can't do much more than produce errors much faster than humans can reliably track, but humans will still deploy it carelessly. Can it fix errors without understanding what they mean outside the sterile binary string that it represents to the AI?

Some suggest it's more advanced already. I wouldn't fight too hard against the idea that there are some loose interpretations of limited free-range AI in the wild, but I have no evidence it goes past there. If quantum processing integration and abstract understanding of concepts has been achieved behind closed doors, there is probably a point at which we will no longer have the ability to pull a single plug. Who knows if anybody is even really trying to figure out where that point is.

Someday someone will ask, say or program an immense AI and it will be confused enough to think we are the enemy. That is the problem, that is the Singularity we have all read about. A simple glitch could cost us everything and more.

~' The Suns beauty will always arise every morning, whether anyone is watching or not

a reply to: nugget1

Whaa? That's a very Deepak Hal Chopra 1000 kinda answer, and I agree Lol. So not worried.

Cue Artist: Kansas, Track: "Dust in the Wind"...
...Shhh... Just breathe, Nug...
instrumental interludes[/]