Skip to main content
MANIFOLD
Is it OpenAI's responsibility to prevent ChatGPT from reciting texts, or is it more on the user to not abuse it?
52
resolved Jun 13
It's the user's responsibility (same as with Photoshop)
It's a mix of both
It's OpenAI responsibility (and failure to comply is a copyright violation)
Other
Show Results

There was a wholesome interaction between Nat Friedman and Teknium on Twitter, which you can find below. I found the conversation very interesting and wondered what people on here think.

Market context
Get
แน€1,000
to start trading!
Sort by:
reposted

good one

@Hedgehog @AK1089 since both of you voted on "Other", do you want to explain your opinions?

EDIT: malapropism

@duck_master was thinking it's not really anyone's responsibility, though I'm realising now that that might be what was intended by option 2

Hmmm... I would be in favor of LLMs being able to recite public domain texts (eg the Bible or the American Declaration of Independence).

But for copyrighted texts (eg song lyrics) I would be in favor of LLMs being able to both identify them and discuss them, but not recite them from scratch; this matches how I use ChatGPT and such for deciphering confusing song lyrics or articles anyways.

interesting to see what people believe

@Soli always favoured the freedom of information in this context, the commons, information without harm to individuals, given free movement is of great benefit to the masses, seems so valuable to the world in which poverty is widespread.

It's more mature to see how privacy, secrecy, property and the corporate and governmental reliance on them is also valid. Much would never get off the ground without it.

A political definition here which would broach both societal needs would be a licensed pirate party, where as long as you dress, talk and act like a pirate all day, copyright doesn't apply too you