Comment on OpenAI says it’s “impossible” to create useful AI models without copyrighted material
pupbiru@aussie.zone 9 months agobranding
okay
the marketing
yup
the plagiarism
woah there! that’s where we disagree… your position is based on the fact that you believe that this is plagiarism - inherently negative
perhaps its best not use loaded language. if we want to have a good faith discussion, it’s best to avoid emotive arguments and language that’s designed to evoke negativity simply by their use, rather than the argument being presented
I happen to be in the intersection of working in the same field, an avid fan of classic Sci-Fi and a writer
its understandable that it’s frustrating, but just because a machine is now able to do a similar job to a human doesn’t make it inherently wrong. it might be useful for you to reframe these developments - it’s not taking away from humans, it’s enabling humans… the less a human has to have skill to get what’s in their head into an expressive medium for someone to consume the better imo! art and creativity shouldn’t be about having an ability - the closer we get to pure expression the better imo!
the less you have to worry about the technicalities of writing, the more you can focus on pure creativity
The point is that the way these models have been trained is unethical. They used material they had no license to use and they’ve admitted that it couldn’t work as well as it does without stealing other people’s work
i’d question why it’s unethical, and also suggest that “stolen” is another emotive term here not meant to further the discussion by rational argument
so, why is it unethical for a machine but not a human to absorb information and create something based on its “experiences”?
Phanatik@kbin.social 9 months ago
First of all, we're not having a debate and this isn't a courtroom so avoid the patronising language.
Second of all, my "belief" on the models' plagiarism is based on technical knowledge of how the models work and not how I think they work.
This would be impressive if it was true. An LLM is not intelligent simply through its appearance of intelligence.
It's a chat bot that's automated Google searches, let's be clear about what this can do. It's taken natural language processing and applied it through an optimisation algorithm to produce human-like responses.
No, I disagree at a fundamental level. Humans need to compete against each other and ourselves to improve. Just because an LLM can write a book for you, doesn't mean you've written a book. You're just lazy. You don't want to put in the work any other writer in existence has done, to mull over their work and consider the emotions and effect they want to have on the reader. To what extent can an LLM replicate the way George RR Martin describes his world without entirely ripping off his work?
If I take a book you wrote from you without buying it or paying you for it, what would you call that?