To be terrified by the rapid progression of Artificial Intelligence?

[FlyRobinFly]

These children in the 2 attached photos are not real, an AI was asked to generate an image of what Lisa and Maggie Simpson would have looked like had The Simpson’s been a live action 90s sitcom rather than a cartoon. Those children do not exist, an AI has generated those faces within a few seconds based on a prompt of a few words. The quality of the image is even just the right amount of distorted to look like the images have come from a 90s live action sitcom. It seems so real…and yet it’s not. Like a window into an alternate dimension or something.

I find it incredibly unsettling.

For me the concern is the artificial creating art and that while community is being very defensive, it has the potential to take over there. There are some very high profile legal cases at the moment because these ai art producing software sites use billions of real copyrighted images to sample from to make the new image. The argument being that the new image breaches the copyright.

I work in design and so have concerns about jobs in the creative industry. I expect most people though this was an area that robots couldn't emulate but we were wrong. Things like graphic design, logo product design, advert production etc will be seriously changed by this.

I'm fortunate that I work in a design field that requires reference to the real world (landscape design) so perhaps I'm safe...for now.

My hobby is making 3d renders of fantasy places and that is an area can be completely decimated by this tech. Privately, individuals decry it. Artstation, a big forum of art creaters has banned ai art but I can see in the future that big companies will start to see it as a cheaper alternative to paying artists for e.g. concept art and it will start to creep in because of market forces. Suddenly an artists skil isn't worth as much and their salary tanks if they can find work at all. There could be major consequences to this.

Crap in crap out that's the crux of the issue for me. Who is in control of this stuff and what are they hoping to gain from it? Will they use it for good?

We know bad people exist and some of them are very clever and some of them are wealthy enough to play around with this stuff.

We don't understand yet how far this could go and yet were messing with it without global ethical codes.

I think it's why Elon musk is so interested in it. He knows how awful this could get. Is he in it for the right reasons and who decides if their reasons are good or bad or right or wrong?

We're on the tip of something, that's for sure.

In terms of replacing humans, I wish it could replace my country's politicians (and quite frankly, this country's too). Imagine PresidentAI-Compassion that calculates levels of poverty, urgent health and social care needs etc and redirects resources proportionally, and that's programmed avoid conflict and to not lie

whilst that's a good concept in principle, @GingerScallop it would always need human interventions to apply the conclusions of the PresidentAI-Compassion's determinations and enable distribution and availability of healthcare and essential resources to the people in need. And don't we know that in poverty-stricken countries with corrupt regimes, no matter how much funding is donated to the right people, it never actually reaches them....

hence why I have zero concerns about AI taking over the world, more like I have increasing concerns about the damage human beings continue to do to each other relentlessly.

The ‘Adele’ song isn’t AI. It was a clever marketing hoax by the singer!

Exactly. I'd never even heard of this. I'm not glued to my phone, so happily going on with my life not being scared or creeped out, it's great 😀

Better in the chat thread if you want a discussion
Putting it in aibu will generate comments/views that are not in line with yours, so if you want a bland discussion, move the thread
And media has been manipulated for years. If you think mass deception is new, you are sadly way behind the curve...

Just to be clear, AI art does not sample from existing art. It’s not taking existing art and chopping it up and putting bits together. What it does is it LOOKS at the existing art and learns its features - eg what colours and shapes are characteristic of a particular artist’s work, it learns that a dog is this particular shape with four legs and two eyes, a tree is this green shape with a trunk, etc. Then when it’s asked to produce a particular object or style it has a very sophisticated understanding of what those things look like and it can generate new variations. NEW variations that follow the same rules but are totally original. It’s very clever.

How's this for AI in the context of a hellish incel shitshow:
twitter.com/heartereum/status/1618973299739951105

None of this is ai. It's machine learning. There is no "intelligence" at play here. It's just a machine trained to do one specific thing. It's just another evolution of a tool.

We are decades away from an actual ai, if ever.

that ChatGPT software which is writing academic essays and passing with flying colours

No it isn't.

I work in a university & this is undoubtedly a concern. But it's not true that it allows students to excel - having tested it, it will produce perfectly passable essays, but limited critical analysis or references that would lead to a higher grade.

Of course, many students are quite happy with that!

Our approach is going to be to educate students on the skills they are losing by doing this & explain the true purpose of research & writing - you won't be able to use ChatGPT in the workplace on a daily basis, for example.

It's a tool. As PP have said, there are both advantages & challenges, as there have been with every new development from the dawn of time.

the something off about them is usually always things like the hands/fingers/ears/nose. There’s a creepy AI Art group on Facebook where at first glance the images look ok but there’s something unsettling about them then you realise that uneasy feeling is coming from the image has weird pointy hands or missing fingers!

The TV show thing that a previous poster mentioned is already here!!
www.theguardian.com/tv-and-radio/2023/jan/09/deep-fake-neighbour-wars-interview-itvx-comedy

The scary part for me is how it’s come out of nowhere.

Er... out of decades of research.

@wherearerhou they know they don't exist because they are images created by a computer program. Other people could look similar but it's not them they don't look the same.

Clearly the OP hasn't ever heard of Alan Turing and his annual Turing Test, which is a computer programming competition to see who can make their AI programme so convincing, the judges think it is a human.

and just to add, Alan Turing was around in the 1950's!

I agree. Unfortunately, as social media has shown us, US tech firms do exactly what the fuck they want to and fuck anyone else.

It's a bit more than that!

"Artificial intelligence (AI) is intelligence—perceiving, synthesizing, and inferring information—demonstrated by machines, as opposed to intelligence displayed by non-human animals and humans."

We're there.

If the book of revelation isn't just a pile of horseshit then I'm beginning to believe that the antichrist isn't a human being after all.

A friend of mine who is a uni lecturer is worried about this. She showed me some software recently and you ask it to write an academic essay on whatever, you can set parameters like word count and it’s there a few minutes later. She says it can’t be picked up plagiarism software but surely if two students used the same software they’d get the same essay? But yes, it was a scarily good response. Though no references!

Sorry, see someone has mentioned the chatgpt software already

There’s literal humans raping and murdering people but you’re scared of AI creating a photo set in the 90s? Weird thing to be outraged about.

AI is not “new” or “surprising”. It’s not even that it’s rapidly progressed in the last few months either - the tool used has likely been in development for years and regularly tweaked before being released for public testing. This didn’t happen overnight, I can assure you this would have been difficult to code.

We are in 2023 and officially in the digital age. Almost everything you access online is collecting data on you, possibly to benefit machine learning. And a lot of the tools we access already, have benefited from machine learning and algorithms. This isn’t the start, we’re already in the thick of it.

DH has been looking at what jobs will exist in a few years. He works in IT and still found it scary.

That's where the shit hits the fan. Not computer generated images.

He was scared of the implications. He will still be employable.

It's likely there will be a backlash before then... And that's scary too.

I think it's the eyes. Really creepy like something out of a horror film.

This image is clearly a different art style. If you asked the AI to make an image that looked like this, it would create something similar to this again.

It wouldn’t suddenly create the same thing but hyper-realistic unless you specifically asked it to do that. These images are not proof of rapid progression at all.

It is all contingent on the terms inputted to the AI, so unless you have proof of the terms used to generate these images then you have no basis for your claim.