this post was submitted on 01 Apr 2025
255 points (97.0% liked)
Technology
70285 readers
3532 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
If you think that AI is only trained on legal images, I can't convince you otherwise.
I mean, you're not giving a very convincing argument.
AI models are trained on the open Internet. Not curated. Open Internet has horrible things.
So is that the Gen AI problem or the open internets problem. It sounds like you hate the open internet and awful people who put real cp online and not Gen AI.
If you're selling a service, I expect you to know where your parts come from and what's in it.
Obviously you don't know what gen AI has become.
What AI are you talking about? Are you suggesting the commercial models from OpenAI are trained using CP? Or just that there are some models out there that were trained using CP? Because yeah, anyone can create a model at home and train it with whatever. But suggesting that OpenAI has a DB of tagged CP is a different story.
Open AI just scours the Internet. 100% chance it's come across someone illegal and horrible. They don't pre-approve its training data.
But you have to describe it. It doesn't just suck in images at random. I imagine someone will remove CP when the images are reviewed. Or do you think they just download all images and add them to the training set without even looking at them?
I think that's exactly what they do. Curation at the quantities that they're working at would require an army.
So you think to train AI you just show it random images without describing what they represent and AI just magically learns? If I then ask AI to create an image of a computer, how does it know what a computer is? Does it just learn this on it's own from all the random images?