Ask HN: Is copyright obsolete in the age of AI?

4 points·blindprogrammer··4 comments

Does copyright become obsolete now that large language models (LLMs) are scraping and ingesting the content we write, and then use it to train themselves? What's the point of being a volunteer content creator for OpenAI or Google if LLMs scrape everything you put online? Why even try?

View on HN

This was the conundrum I faced when I learned that everything was being scraped to train LLMs. I haven't figured out a reasonable solution to this at all and was forced to remove my websites from being accessible to random web visitors.

I'm very eager to find a way that I can make my sites accessible again.

Maybe requiring credit card payment would filter out LLMs. Is this gonna lead to closed-off paid websites ?

That might be effective, but requiring credit card payments is an unacceptable solution to me.

First, it would introduce a commercial aspect to my websites that I don't want them to have (too much of the web is commercial as it is).

Second, it would require me to have some sort of merchant account just to have a website. I don't want that additional hassle.

In the end, it doesn't appear to be a better solution than to just make my websites invite-only.

Well there is some flag that you can set in robots.txt of your website. But however I am sure, not many AI vendors respect this flag.

Which is why Google Gemini doesn't works to that level of magical greatness, I guess. They go with the ethical route of data collection.

Reply on news.ycombinator.com