alt Hacker NewsI mean the models have literally trained on:
1. Child porn
2. Stolen music
3. Private github repos, before that was 'stopped'
4. Illegally pirated books
Them training on company prompts against the terms of service would be one of the least bad things that these companies have trained AI models on
Why do you think a company - willing to break the law for child porn - won't break the law when it comes to your personal data?
These folks need to feel repercussions so hard their souls flee to the afterlife leaving only their sad, dead husks behind.