105
submitted 2 months ago by ooli3@sopuli.xyz to c/coolguides@lemmy.ca
you are viewing a single comment's thread
view the rest of the comments
[-] IntriguedIceberg@lemmy.world 2 points 2 months ago* (last edited 2 months ago)

But then we have a problem with AI's process, not the underlying technology. What I mean to say is, if I were to train and run my own model in an isolated local computer, using ethically sourced* training material, then would it still be bad? Like, are we mad at the means needed to make the technology work or simply at the technology? Because if it's the former, then we can do something about it. Agriculture is a hell of a technology, but there are very detrimental ways to apply it.

*Let's gloss over what ethically sourced really means for the sake of discussion.

[-] GreenBeard@lemmy.ca 3 points 2 months ago

Sure, if you built an AI on your own machine, trained it entirely on public commons and voluntarily obtained data with the active consent, and powered it entirely on solar power and wind turbines, to do jobs without intrinsic value to human development, people would have a lot fewer objections to it. But you didn't. And you won't, because it would take resources that exceed anything you have available to do so. Much like genetic modification, there are motives and methods that potentially have real value, but they don't tend to have significant return on investment and so are simply not done, and what is done ranges from suspect to objectively exploitative. You cannot create an ethical AI in the current environment, if such a thing is even possible.

[-] Darkassassin07@lemmy.ca 2 points 2 months ago

It may very well be the former; but I don't believe 'ethically trained' and 'useful' is an achievable combination in AI models, particularly within our capitalist world/society.

Even if such a thing was created; it'd be impossible to verify and distinguish among the flood of media produced by those that just don't care about the broader picture and those effected. The detrimental effects of global AI infrastructure has greatly overshadowed the potential for 'ethical' AI implementation; and now we just don't have the means to sort through 'ethical' vs not, leaving no option but to reject it outright.

[-] wieson@feddit.org 1 points 2 months ago

Do it then. Start by describing ten thousand images, so your local ai can learn how to recreate them.

Ten thousand isn't nearly enough btw


I've worked and researched with ai and trained a model. I like classificators and CNNs, I don't like chatbots and image generators.

[-] MrQuallzin@lemmy.world 0 points 2 months ago

You can't train your own though. You can't make your own AI that knows what a cat or Picasso is without it using works that are not yours. You'd never be able to tell it to write you a novel because it wouldn't have any to work with.

[-] IntriguedIceberg@lemmy.world 1 points 2 months ago

But if I write a novel, am I not also taking into account every novel I've ever read? Can you draw something without first looking at other drawings?

[-] MrQuallzin@lemmy.world 2 points 2 months ago

AI is not sentient and cannot learn or read. It regurgitates a "best guess" and knows nothing. If you use drawing tools to make shapes in Paint, congratulations! You made art, not the computer. If you write the shittiest fanfic in the world, YOU made it, YOU adapted your experiences into your own work.

AI does not and never will do it. Stop defending it.

[-] IntriguedIceberg@lemmy.world -1 points 2 months ago

You moved the goal post. I thought the problem was using other things as references for the creation process.

[-] MrQuallzin@lemmy.world 2 points 2 months ago

Why are you fighting so hard to defend AI slop? There was never an issue with anyone using references. The issue is stealing and regurgitating other people's works and destroying the environment in the process. I'm sorry if that is too advanced for you, as I'm not the first to tell you this.

[-] IntriguedIceberg@lemmy.world 1 points 2 months ago

When the ad hominem attacks start, it's time to drop it. Than you for the civil discussion while it lasted.

[-] MrQuallzin@lemmy.world 3 points 2 months ago

Apologies, I'm trying to do this while also feeding a very fussy newborn.

I'll instead be an adult and say that I don't want to continue this discussion. I have better things to do, and you're getting pedantic about what defines "art". It's a lose-lose situation.

[-] IntriguedIceberg@lemmy.world 2 points 2 months ago

Agree to disagree. Thanks for being civil. Best of luck with your newborn.

this post was submitted on 04 Feb 2026
105 points (68.7% liked)

Cool Guides

6434 readers
104 users here now

Rules for Posting Guides on Our Community

1. Defining a Guide Guides are comprehensive reference materials, how-tos, or comparison tables. A guide must be well-organized both in content and layout. Information should be easily accessible without unnecessary navigation. Guides can include flowcharts, step-by-step instructions, or visual references that compare different elements side by side.

2. Infographic Guidelines Infographics are permitted if they are educational and informative. They should aim to convey complex information visually and clearly. However, infographics that primarily serve as visual essays without structured guidance will be subject to removal.

3. Grey Area Moderators may use discretion when deciding to remove posts. If in doubt, message us or use downvotes for content you find inappropriate.

4. Source Attribution If you know the original source of a guide, share it in the comments to credit the creators.

5. Diverse Content To keep our community engaging, avoid saturating the feed with similar topics. Excessive posts on a single topic may be moderated to maintain diversity.

6. Verify in Comments Always check the comments for additional insights or corrections. Moderators rely on community expertise for accuracy.

Community Guidelines

By following these rules, we can maintain a diverse and informative community. If you have any questions or concerns, feel free to reach out to the moderators. Thank you for contributing responsibly!

founded 2 years ago
MODERATORS