Oh sweet baby Jesus no, burn AI to the ground, humanity can't be trusted with it

Recommended Videos

Bob_McMillan

Elite Member
Aug 28, 2014
5,694
2,256
118
Country
Philippines

Uh-huh. Same old, I guess, with new toys.
I saw a similar situation where a police department in the US used AI to literally do a "ENHANCE!" meme on grainy ass CCTV footage of a suspect, then posted the AI slop on Facebook telling citizens to call in with any information. The suspect was a black dude, and the AI very clearly gave him features he did not have (such as a completely different and dare I say more stereotypical hairline). When the cops were called out on it, of course they got all pissy and defensive.

Back home they do this all the fucking time, albeit they keep it old school with photoshop (see an amazing example below). Although I wouldn't be surprised if they've "upgraded" to AI.
emwe1liucaekkq3.jpeg
 

Drathnoxis

I love the smell of card games in the morning
Legacy
Sep 23, 2010
6,521
2,470
118
Just off-screen
Country
Canada
Gender
Male
I saw a similar situation where a police department in the US used AI to literally do a "ENHANCE!" meme on grainy ass CCTV footage of a suspect, then posted the AI slop on Facebook telling citizens to call in with any information. The suspect was a black dude, and the AI very clearly gave him features he did not have (such as a completely different and dare I say more stereotypical hairline). When the cops were called out on it, of course they got all pissy and defensive.

Back home they do this all the fucking time, albeit they keep it old school with photoshop (see an amazing example below). Although I wouldn't be surprised if they've "upgraded" to AI.
Was this the suspect?
 

Agema

Overhead a rainbow appears... in black and white
Legacy
Mar 3, 2009
10,996
7,953
118

Agema

Overhead a rainbow appears... in black and white
Legacy
Mar 3, 2009
10,996
7,953
118
I didn't mean that I expected to be taught how to use AI (that was explicitly not a part of the course, which is why I signed up for it haha), but the entire course is based on the assumption that AI is useful, does make people productive, etc. I expected the course to dedicate some time to prove this assumption, especially since AI is plagued by hype. Not once did someone actually back up the claim that AI makes us more productive, we just went straight into the necessities of creating legislation to encourage AI development and tech sovereignty, blah, blah.

So an AI skeptic such as myself is just left wondering while I should care about AI, besides the threat that it is to my career. The sole lecturer that shared about practical implementation of AI talked about using machine vision to stop suicide attempts in the subway and to optimize shipping packaging. But neither of these examples entail PC parts costing an arm and a leg or data centers ruining communities. I feel like the course failed to define what AI even is.
Okay, yes, I see that.

I think the problem is that courses need to teach something, but they don't know what.

So, imagine that AI as a tool can if used well improve productivity, but if used impair it. At the same time, organisations have systems that work: they may be better or worse, but they are generally known and stress-tested. AI may often disrupt these, causing a fundamental loss of productivity by damaging established processes, even if it creates productivity gains elsewhere. It's going to take lots of time - potentially years - to work out how to best use AI to boost productivity and minimise damage and disruption. Secondly, the economics of AI are utterly busted. It's subsidised up the wazoo, so no-one's paying actual cost. Once AI companies are forced to bill customers properly, we could see AI use dramatically shrink, because it turns out some of the AI productivity gains that were established were actually cost-ineffective.

And this is I think where lots of courses on AI are. What is it going to be useful for in the long run? They don't know. The best that they can do is say some speculative ideas and hope some of them pan out. As for its optimism, I guess it made the assumption that anyone who was going to sign up wanted (or thought they would be forced) to use it.
 

Xprimentyl

Made you look...
Legacy
Aug 13, 2011
7,589
5,956
118
Country
United States
Gender
Male
I already posted this video in the "Some Good Music" thread, but I'm posting it here for a different reason. I've watched a bunch of reaction videos, and the one sticking point that turns people on or off is the use of AI.

The visuals and vocals are AI, but the lyrics were written by a human, Dallas XY. I've heard it took them a very long time (over 300 attempts is what I've read) to feed the AI tool the inputs necessary to produce what we see here, and I think is one of the better uses of AI that doesn't threaten the human creative process. Some detractors dismiss it because it's "not real," but to that I say it's just as real as the Talkbox made famous by Peter Frampton, or artists like Near The Parenthesis who composes beautiful piano arrangements using production software that would require more than two hands if one were to try and play them on an actual piano. This is easily one of the better uses of AI as a tool that doesn't replace/discount the human effort. If AI stayed in this lane, I'd not be against it, but if it ever starts writing brilliant bars like this, Skynet is inevitable.