Support NeoGAF

[s_mirage]

I don't see any problem whatsoever. AI is a tool which can enhance productivity. If it can be used in a way that facilitates the production of a higher quality product, have at it. Now, it can also be used to produce shit, but people don't need AI for that.

I'm also not in favour of management enforcing its use simply because it's seen as the new hotness; I've seen that with non-AI IT related things before.

I know someone who hates ai because it "steals" art work.
This is the same person that traces other people's art and style for his own use.
Like bro....

Reminds me of that samdoesart guy. Complains about people copying his style with AI (despite style not being copyrightable) while a bunch of the works that he was selling appeared to be uncredited and unauthorised draw overs of photographs and film stills.

It's also similar to the multitude of anti-ai whiners constantly going, "whaaa, AI art is stealing", while simultaneously using pirated software and denying or trying to excuse their own hypocrisy.

"B, but pirating software only hurts corporations", they say. And what about the programmers, designers, artists, etc, who were all employed to create that software? What happens to them when the software underperforms thanks to everyone pirating it? What about small developers whose work is pirated?

 

[BlackTron]

It's okay to use GPS daily, it becomes bad if you use it on a known route and follow it blindly into a ditch

 

[FukuDaruma]

the big problem with AI generator images is that it is ripping off everyone. I tried an image of a warrior guy the other day and he had batman gloves and looked like John cena.

Whaaaat? no way! all they do is take "inspiration"...

Generative AI Has a Visual Plagiarism Problem

Experiments with Midjourney and DALL-E 3 show a copyright minefield

spectrum.ieee.org

«Midjourney apparently sought to suppress our findings, banning one of this story's authors after he reported his first results.»

 

[FukuDaruma]

I don't see any problem whatsoever. AI is a tool which can enhance productivity. If it can be used in a way that facilitates the production of a higher quality product, have at it. Now, it can also be used to produce shit, but people don't need AI for that.

I'm also not in favour of management enforcing its use simply because it's seen as the new hotness; I've seen that with non-AI IT related things before.



Reminds me of that samdoesart guy. Complains about people copying his style with AI (despite style not being copyrightable) while a bunch of the works that he was selling appeared to be uncredited and unauthorised draw overs of photographs and film stills.

It's also similar to the multitude of anti-ai whiners constantly going, "whaaa, AI art is stealing", while simultaneously using pirated software and denying or trying to excuse their own hypocrisy.

"B, but pirating software only hurts corporations", they say. And what about the programmers, designers, artists, etc, who were all employed to create that software? What happens to them when the software underperforms thanks to everyone pirating it? What about small developers whose work is pirated?

The bottom line is that generative AI is a deeply extractive and unsustainable business model, based on the precariousness and massive plundering of people's creative work to enrich large corporations, while masquerading as a revolutionary advance, a supposed "democratization" of art, or simple "inspiration" comparable to that of humans.

AI not only exploits cultural heritage—such as attempting to replicate the unique style of Studio Ghibli—but also undermines creativity by prioritizing economic profits over authenticity and the human value of the artistic process.

Piracy as individual access without profit motive is not the same as large-scale piracy for profit, which is what would really be comparable to generative AI, and which has always been punished legally with thousands of harsh sentences.

Piracy is illegal, but it does not alter or reuse the work. If you pirate a Ghibli movie, you watch it as is, without modifying it. At most, you reduce the creators' income, but you don't take away their authorship or transform their work, and above all, you don't get rich doing it.

Generative AI systematically exploits the work of others for corporate profit.

The problem is not people making little drawings with AI, just as the problem has never been people downloading an album or Windows or Photoshop. The problem is large corporations making billions by plundering and professionally destroying millions of small professionals and companies around the world, stealing their work because legislation always arrives too late and these same corporations are already working to ensure that it doesn't exist by buying off legislators.

 

[yogaflame]

I don't see any issues with AI. But for serious decision making, it will always be human intervention. AI is just for support and checking and verification.

 

[FukuDaruma]

I don't see any issues with AI. But for serious decision making, it will always be human intervention. AI is just for support and checking and verification.

An AI Mistook a Doritos Bag for a Gun and Called the Cops on a Teenager

An AI-powered gun detection system mistook a bag of Doritos chips as a weapon — and called the cops on a 16-year-old student.

futurism.com

 

[CuNi]

AI has no Soul

Which is an argument for what exactly?
There is no scientific proof in a soul, science in general doesn't believe in a soul and neither do I.
So to me, all Art is "soulless".

I never understood the argument that AI creations are worse simply by the fact that they are made by AI instead of a human being.
That's to me wrong and incorrect on so many levels.

 

[ahtlas7]

Nothing wrong with it. Keep training your replacement. Cheers!

 

[belteshazzar]

Like any tool, it can be abused. I'm a developer and use AI to automate tedious tasks (scaffolding, etc), to help me learn something new and sometimes to optimize or help me debug. The important thing is I ALWAYS double check to make sure the LLM is not hallucinating, just plain wrong or missing important stuff, which often happens. The problems arise when lazy developers don't bother with this last crucial step or over rely on AI too much. It has already happened where I work. When code review time comes around and we find stuff that rises some eyebrows and then their excuse is "dunno, that's what the LLM told me" is, of course, inexcusable. You cannot shift the blame to the LLM. As a rule of thumb, if I ever use anything an LLM gives me I need to understand every single line its giving me. Otherwise, you're just asking for trouble.

 

[Valonquar]

Here is why it is bad:

 

[Signalis]

AI probably has it's uses. But I won't be a GPTard and use it for everything like a lot of people do. Because a lot of people have lost the ability to think for themselves and use LLMs for everything.

Especially when a thread on NeoGAF starts with: "So I asked ChatGPT....". Then it's easy to know that you're not dealing with someone who can actually think for themselves

 

[Draugoth]

The bottom line is that generative AI is a deeply extractive and unsustainable business model, based on the precariousness and massive plundering of people's creative work to enrich large corporations, while masquerading as a revolutionary advance, a supposed "democratization" of art, or simple "inspiration" comparable to that of humans.

AI not only exploits cultural heritage—such as attempting to replicate the unique style of Studio Ghibli—but also undermines creativity by prioritizing economic profits over authenticity and the human value of the artistic process.

Piracy as individual access without profit motive is not the same as large-scale piracy for profit, which is what would really be comparable to generative AI, and which has always been punished legally with thousands of harsh sentences.

Piracy is illegal, but it does not alter or reuse the work. If you pirate a Ghibli movie, you watch it as is, without modifying it. At most, you reduce the creators' income, but you don't take away their authorship or transform their work, and above all, you don't get rich doing it.

Generative AI systematically exploits the work of others for corporate profit.

The problem is not people making little drawings with AI, just as the problem has never been people downloading an album or Windows or Photoshop. The problem is large corporations making billions by plundering and professionally destroying millions of small professionals and companies around the world, stealing their work because legislation always arrives too late and these same corporations are already working to ensure that it doesn't exist by buying off legislators.

AI is closer to theft than actual piracy (sharing some company content's for free on the interwebs) Disney and Warner Bros touch it up with dozens of examples on their lawsuit If you ask Midjourney to 'generate me a yellow cartoon character 'chances are the Simpsons will show up people also forget they charge subscriptions for their models.

You also has to consider: If these models need training data to generate children pictures and videos but are trained on the entire internet, where is the training data coming from? If you ask OpenAI CFO if they trained their models using Youtube database she will not answer you.

There's nothing ethical about Generative AI.

 

[Ainem Enamas]

Which is an argument for what exactly?
There is no scientific proof in a soul, science in general doesn't believe in a soul and neither do I.
So to me, all Art is "soulless".

I never understood the argument that AI creations are worse simply by the fact that they are made by AI instead of a human being.
That's to me wrong and incorrect on so many levels.

Because you simply don't know
This Reality is so much fucking more

And if the dev-time is reduced by years, then it's completely fine with me.

 

[hussar16]

Calling Generative AI creative is a stretch. The tech would be probably worthless without the internet to scrap data for training. It's basically thousands of GPUs lined up running extremely inefficient predictive algorithms though a black-box not so different of a slot-machine. You don't know how the model came to the solution of your output neither if the results are correct. Even with local models using workflows you are still limited in what you can achieve.

The monkey compassion is good, Machine Learning is essentially Infinite Monkey Theorem on steroids. You can teach a monkey how to paint but it doesn't mean it knows what the fuck it's even doing, it's basically just mimicking the humans, the same goes for LLMs, which don't really think, they just can process a lot of information.

The artist still has control of every single pixel with traditional tools, which is why Generative AI still hasn't been massively used in movies/anime/games industry as they are still behind traditional tools in results.

Now the day AGI happens than it's pretty much game over to the creative field, but most experts agree we won't be living to see it.

the thing i have with this is most art and art styles have already all ben done before so artist have already ben copying each other and art styles for decades so whats the problem if a computer does it. and honestly for the talk how creative humans are tht goes only for a very few who make unique things. most humans just copy and do the same as ai does. the issue i have with it is the algorith breaks to much and it is not reliable and not unique enough yet.

 

[yogaflame]

An AI Mistook a Doritos Bag for a Gun and Called the Cops on a Teenager

An AI-powered gun detection system mistook a bag of Doritos chips as a weapon — and called the cops on a 16-year-old student.

futurism.com

Haha! That is why the final decision should always be human. I see that AI can be useful as support only and for double checking and quick calculations.

Reminds me of the droids in Star wars, R2D2 and C3PO. Both are very useful for support, translations, quick calculations, information, and repairs ships, but humans are still the final decision and will be the one to be followed.

To avoid ethical danger there should be basic ethical programing laws in place and safeguard in its memory chips. Reminds me of Asimov. This is what UN must be aware of and discuss in the future on guidelines on AI and robots. And AI must not be treated as a human regardless of what it can do and accomplish especially it is soulless. It has to have limitations and safety switch. It is just a support tool, and nothing else. There will be no emotional attachment to it.

 

[CuNi]

Because you simply don't know
This Reality is so much fucking more

To be fair, no one knows, but saying "AI has no soul" implicates that someone does know, which simply is wrong.
And I agree!

Living should be way more than it currently is.
This will sound socialist, but I really think peak human living conditions would be with UBI and everyone being able to do what he wants to do.

 

[Generic]

I dislike AI because it's popular.