[X] My Assistant

I recognize the potential of AI, but I feel it’s not truly ready until I have the chance to use it myself. Aside from chatbots, which significantly reduce the learning curve for almost everything, AI drawing—like the IllustriousXL model (a Sable Diffusion XL model)—has genuinely impressed me. It’s far more advanced than AI from two years ago, like SD 1.5 or NovelAI v1. Today's AI is remarkably powerful.

My opinion is quite simple.
Just like technologies like Google, it will inevitably create a dependency for our society and future generations.
We have to be able to have AI used as a tool of research and assistance but not for reliance.
There's a sense of moderation needed.

(IMG:[i.imgflip.com] https://i.imgflip.com/7uji6v.jpg)

I despise it. Apart from solving complex algorithms and running simulations, it's a useless, broken tool that humany is better off without. Generative AI is especially useless because it plagiarizes actual artists to make fugly images no one in their right mind would think it's good art. Fuck AI -.-

AI will usher more smooth-brained users, consumers, instant-gratifying people, and more exploitation. Only few things will be left untouched by AI. The world needs a global reset, and it will be done partly by AI.

(IMG:[invalid] style_emoticons/default/blink.gif) (IMG:[invalid] style_emoticons/default/laugh.gif) ehhh idc

I am against "AI" without a shadow of doubt.

"AI" is the latest mostly useless tech trend similar to crypto. There are lots of confusion regarding what even "AI" is right now, So it's basically being sold as this magic that will grow exponentially and take over everything while others compromise by saying it will be another tool to be used by professionals for the job. Lots of fear mongering too, naturally demoralizing some people out of their passion. Already a net harm to several communities.

There are lots of tech startups desperate for a slice of the clueless investor pie. "AI" developers are trying to shove their toys into everything they can think of. And for a good reason. Some of them make it big and make millions upon millions in maybe just a few years while not even delivering something applicable in real life, or at least something sort of useful. These people are motivated by money only and could not care what harm or good their actions bring as long as the number in their bank account goes up. I am overall not fond of modern investment gambling and tech industry due to their overall harmful on legitimate people.

AI has no practical use. Art prompted by "AI" looks ugly. Voice generated sounds odd. You can't trust them for banking, laws or other important things either. A common defense against this is "AI" will get better, but it won't. It's almost like machines can't learn like humans do. And the humans already do the job of learning, further proving AI has no use in a world with billions of humans up for the task. Machines are better off being calculators and doing if else stuff which are coded deliberately by a human. In that case they are actually very useful and revolutionary. If you want something that can exponentially grow and take over everything, you are better of investing in genetic engineering in hopes of creating trans super humans with great intellect instead of "AI".

What's more I didn't even talk about morals yet. So far I only talked about cold hard facts regarding "AI" revolution's effects. The models are trained using stolen assets. Yet they are for profit business. Someone stealing your data and using it to make profit without even asking you, this is plain immoral.

And the potential long term effects. "AI" may be deliberately abused to spam the internet more efficiently. There are real organizations funded by governments tasked with spamming internet right now. They alone can take down entire sites often forcing site administrators to range ban, perhaps also banning legitimate users. Could "AI" be used to spam the internet with illegal content harder? Dodge captchas constantly? Eventually giving enough excuse for governments to enforce digital ID laws to "protect" us from their own "AI" spamming harmful content? Perhaps... Perhaps not...

Besides we have plenty of humans to deliver us any content we want. Why even engage with second hand abominations produced by "AI" when I can directly consume their source material: Human content.

There's one reasonable anti-AI opinion, and it's that it's still too low-quality to be very useful. Consistency is a solved problem if you know what you're doing, I really mean just low quality, not inconsistent.

Every other anti-AI opinion is basically wrong. "Misinformation", "deepfake porn", "muh environment", "muh capitalism" are all cringe boomerslop. They'll die out soon. "It will take jobs" is good if it happens, but we still use paper documents and fax machines, it won't happen fast enough to matter, much more important technological developments will eclipse generative AI in the near future. "It's stealing IP" rests on a complete misunderstanding of how it works.

I beg people to maybe try extrapolating trends to 2030 or even 2026, instead of assuming it will stay at its current state.

I hope that Trump will put tariffs on AI.

And I hope AI will tell us how to get rid of Trump.

Or perhaps Trump is actually an AI.

So far because I haven't been affected by it, more good than bad

AI really is a good instrument for anything EXCEPT art. AI can translate better than google, it does great work on projects and stuff like that, but it will never create real art. No matter how this "art" will look, it will never be real without a soul that can feel pain and joy creating it. The worst part is that AI is too much forced by corporations nowadays.

I really don't understand people who create images with AI and call this art. I’d rather have AI handle the laundry and dishes while I create art, not the other way around.

I don't know why people always talk in terms of black and white, good or bad, and so. AI like other inventions, are tools, objects, which will be dependent on how you use it.

IA has the potential to save us a lot of time on tasks where all the important variables are easily recognizable, but this could take a lot of jobs (it will surely do it with some of them) as robots did (assembly lines) some years ago.

I think only few companies will successfully on the right way and I mean "right" by not taking extreme decisions, fire only the unnecessary jobs, save time on simple tasks or give a hand to their employees through IA. On the other hand, most of the companies will try to earn more money greedily by firing as much as they can to save money; giving a terrible customer service when they are asked for something out of the norm which AI may not satisfy successfully, and keep doing that as long as they get more money.

Regarding to artistic AI art, it should be separated from human art. AI can generate a lot images, but somehow for me it still has some mistakes and although those images are different (character, position, background, etc.) those feels like the same (like barbies just with different skin color and costume). It's easy to get bored, galleries of this type of art should be short.

"Not the AI we needed, but the AI we deserve."

When photo cameras improved, it probably took a lot of painters out of jobs. Only really the good ones remain.
When TV reached everyone, radio was slowly dying. I still remember people in my village listening to football commentators on the radio. Only audio is crazy for sports that are better to watch visually. The TV signal was weak back then. The only entertainment was radio.

And today, most young people flock to social media: YouTube, TikTok, etc. Most TV stations struggle to attract new audiences because they don't have the freedom that private content creators do.

AI is a tool. It is definitely going to kill some jobs, but other opportunities will surely appear.

They're good to deal with things which cannot be done with computers before. But need to control the expectations.

on certain cases, it is gonna be useful, it is like removing the background noises and refining a voice recording in an average room, or generating basic sketches to enhance efficiency on producing animations.

but we shouldn't be expecting AI to do everything, especially it is generally just a software of billions if not trillion lines of codes that could utterly fail due to one fatal error.

My opinion is quite simple.
Just like technologies like Google, it will inevitably create a dependency for our society and future generations.
We have to be able to have AI used as a tool of research and assistance but not for reliance.
There's a sense of moderation needed.

The problem is that Google at the end of the day is effectively a communication tool. It's a service that connects people seeking information with people providing information. That is a gross oversimplification, but it's only really a middle man.

AI in its current LLM-based implementation will never be more than a pattern-recognition and prediction algorithm with tons of data inserted into it. That's its fundamental mode of function, so it isn't some kink to iron out as the technology develops. You can try to manually control its outputs, but at that point the whole "AI" part is just an unnecessary and unreliable addition to a process that already existed since the invention of search engines and websites, and it would require unrealistic manpower for it to be this curated all-around information source people want it to be (and are already using it as, sadly). Any LLM-based "AI" will continue to spit out random shit with no regard to its own correctness or any deeper insight, or really even a great way to check why it's saying the things it is without intensive case-by-case backwards engineering. If someone is spreading falsehoods or straight up factual errors through a website indexed by Google, it's fairly effortless to find out who or what's to blame. The "AI" doesn't even know it's doing it, and it's difficult or even impossible to prevent it from making a similar mistake again without the aforementioned manual case-by-case curation. A horrible thing to rely on.

AI has no practical use. Art prompted by "AI" looks ugly. Voice generated sounds odd. You can't trust them for banking, laws or other important things either. A common defense against this is "AI" will get better, but it won't. It's almost like machines can't learn like humans do. And the humans already do the job of learning, further proving AI has no use in a world with billions of humans up for the task. Machines are better off being calculators and doing if else stuff which are coded deliberately by a human. In that case they are actually very useful and revolutionary. If you want something that can exponentially grow and take over everything, you are better of investing in genetic engineering in hopes of creating trans super humans with great intellect instead of "AI".

AI has the "practical use" of being sorta-kinda correct about 90% of the time, while being really cheap to utilize. If there's anythng the modern neoliberal world order has shown, it is a preference for doing things poorly to save a buck in payroll. So unfortunately AI will likely be around for a long time even if it doesn't improve, although the hype may die down a little bit.

Every other anti-AI opinion is basically wrong. "Misinformation", "deepfake porn", "muh environment", "muh capitalism" are all cringe boomerslop. They'll die out soon.

There is quite literally nothing more "slop" than anything touched by AI. A good measure of this is to ask AI about something complicated you're already very familiar with. Watch the AI be wrong, then ask it the same question a number of times and maybe, MAYBE, on the 5th try it will be correct. Of course since this is a topic you're already familiar with you'll recognize the correct answer, but if you're using it for something you don't already know the answer to before asking, you have no way of telling, negatings its entire function. An "information source" that may or may not be correct and has no way of differentiating between the two is completely useless.

ChatGPT still can't even handle spreadsheets or simple coding tasks (anything more complex than straight up copy-and-paste from online tutorials) without making basic mistakes, and those are the types of problems it ought to be the best at. I've had ChatGPT and other "AI assistants" respond countless times with code that doesn't even compile, or data that I could see with a naked eye is incorrect, and each time they do they provide the answer with an air of confidence, not having the human insight that it may not know the answer. Since I am a boomer I actually had to take the time to learn things on my own at some point, so I recognize these mistakes. A generation that effectively skips the process of developing their own skills due to reliance on AI won't. Typing a question into ChatGPT and copy-pasting the answer is not a skill nor will it teach you how to process information or make your own inferences.

Recently, when I searched some galleries, there were always too many [AI-generated] in the results, but I want to see more human-made drawings. I need to filter the AI tag.

There's nothing I can say here that hasn't been said already by others in this thread, but I do feel like as a flesh and bones artist myself I should slide in my two cents;

(My thoughts of AI for stuff like manual labour inventions is still something I'm working on, so I'm mainly talking about AI art here.)

Aside from all the points made about generative AI's inherent issues, one of my most glaring issues aside from it that I don't think I see enough is how the mere usage of it is an insult to artists who have worked hard on their craft. Think of your favourite hentai artist, most nsfw artists/artists in general have started way before AI because art has been a passion for millennia. It's not like stuff in tech or other job opportunities that would be considered more modern in comparison to art itself wherein you could pick up the skill in just some months and you're set for the job. Being good at art, especially to the degree that many artists are, takes years. Learning curves are different from person to person, always. To grab an AI model, just slam in keywords, maybe even struggle to get it to generate the right amount of detail for a few more minutes, then say that you're as much of an artist as any other artist is just a spit in the face. It's like getting your parents to do all of your schoolwork and then saying that you're the best student in class. If you use AI, depend on AI, you simply have not/do not put in the same amount of effort as artists who draw day to day, month to month, hell even season to season just to maintain and improve their skills.

Editing does not account for that either. If an AI "artist" struggles with editing or getting the right keywords, then that is their own fault. It is not part of the trials and tribulations, it is a problem they have brought onto themselves by choosing to depend on such a soulless tactic rather than simply taking the time to learn art themselves. If you can edit an AI-generated photo, then what's stopping you from just learning? If AI is as difficult as those artists try to make it seem, then they're better off just doing manual studies upon studies of their preferred art style. Because it's not just about the politics, ethics, and whatnot that has already been repeated over and over, it's about dignity and tenacity, it's about character, passion, integrity, and all these other big words because that's what art is—big words that are not spoken but instead funneled through visual beauty, that challenges the sight meaningfully. This applies to all artists, even the artist with the most grossest fetish in your opinion has worked their ass off to get to the point where they are at. They didn't turn to AI for any cheap shortcuts because they wanted to make art, they're that passionate about their fetish that they took the struggle head on. All art has meaning, if it isn't political then it has meaning. Actually doing analysis of nsfw works would be arbitrary because of course you're there to goon not to write an ethics essay on why flat is justice, but the underlying premise is there; AI art cannot, and can never measure up to the standards that already existing hentai artists have placed.

recently I've seen some kids both online and off that have started using the word ''generate'' in lieu of create or draw. Like they would ask ''can you generate more bluey'' or ''generate a mermaid!"

How did that meme go? Makes me wish not to live on this planet anymore.

This post has been edited by Drone-chan: May 23 2025, 01:05