General Discussion

intellectual property is ongoing.

And the AI companies' bots stealing from websites are doing an insane amount of scraping, putting some websites at risk of going out of business. See this thread

https://www.democraticunderground.com/100219421175

about how often they scrape some sites. OpenAI scraping one site "hundreds of times a second.". Anthropic scraping another "almost a million times in a 24-hour period."

Of course these AI companies don't care if they put websites they steal data from out of business, as long as they can steal all that intellectual property first.

You wrote that you "don't use ChatGPT for thinking, writing or art."

I forgot to mention coding, another common use, though LLMs hallucinate while coding, too, and those errors aren't all caught.

But it really doesn't matter what you use ChatGPT or any other generative AI tool for. They are all trained unethically and illegally.

OpenAI has admitted in court filings that their AI tools won't work if training data is limited to what's in the public domain. Licensing is expensive, and although they've made token efforts at licensing some of the content they use, it's mostly for PR so they can try to claim they didn't steal everything. And they already stole as much of the world's intellectual property as they could get, which is why the AI companies are fighting all attempts to make them reveal what's in their training data. OpenAI even told one court that they oh-so-conveniently "lost" some of the info on training data.

They're crooks. Thieves. 21st century robber barons.