OpenAI insists it can't sufficiently train AI models without copyrighted material
The leading company in AI technology says public domain material is not enough to properly train its models.
OpenAI has become synonymous with the massive boom of artificially intelligent technology over the past few years. Among the many common criticisms for AI is its use of copyrighted materials without consent from the original owner. It’s a privilege that OpenAI is now defending, stating that the company’s business model relies on it having access to copyrighted material.
OpenAI filed to the House of Lords subcommittee where it defended its right to train its AI models on copyrighted material, as reported by The Telegraph. "Because copyright today covers virtually every sort of human expression — including blog posts, photographs, forum posts, scraps of software code, and government documents — it would be impossible to train today's leading AI models without using copyrighted materials," the statement reads.
Source: OpenAI
OpenAI went on to say that restricting its usable material to older content that has gone into the public domain would be “an interesting experiment,” but not sufficient for the needs of a modern language model. The company also argued that there is no law forbidding the use of copyright material for training AI models.
There’s been a lot of debate surrounding the use of copyrighted artwork, writing, and other materials to train AI language models, so it’ll be interesting to see how the matter is handled legally. For more news in the world of AI, stick with Shacknews.
-
Shacknews
replyDonovan Erskine posted a new article, OpenAI insists it can't sufficiently train AI models without copyrighted material
-
OpenAI needs to answer why it should be legal for an Etsy store to make Disney content
-
you mean like all these? https://www.etsy.com/search?q=disney&ref=search_bar
-
Yeah that's not legal.
-
-
why is OpenAI responsible for Etsy seller's adherence to copyright law? You could find infringing stuff on Etsy long before AI existed
-
That's a straw man question if I've ever heard one. The point is not "Etsy's infringement does it exist" but rather "it's not legal now, so why SHOULD it be legal?"
-
I literally don't understand why OpenAI needs to answer this question. Does Adobe also need to answer it to justify why their software should be legal since it's enabling infringement?
-
OpenAI's software is built on infringement. All their training sets are comprised of data they've systematically copied from copyrighted content. You already know that.
"We can't make money if we can't wantonly copy everybody's stuff" is not a good defense for copying everybodys stuff. "OpenAI should be allowed to do this because AI is great" is not a good defense either.
I know your whole schtick is vehemently defending major corporations, but it'd be nice if you stopped being purposefully obtuse. You know exactly what the objection to OpenAI's - and every other generative AI company's - business model is. -
"In order to make money AI must copy these works"
"in order to make money my Etsy shop must copy Disney"
-
-
-
-
