Technology

37737 readers

581 users here now

A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.

Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.

Subcommunities on Beehaw:

This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.

founded 2 years ago

MODERATORS

[email protected]

251

OpenAI says it’s “impossible” to create useful AI models without copyrighted material (arstechnica.com)

submitted 10 months ago by [email protected] to c/[email protected]

245 comments fedilink hide all child comments

Apparently, stealing other people's work to create product for money is now "fair use" as according to OpenAI because they are "innovating" (stealing). Yeah. Move fast and break things, huh?

"Because copyright today covers virtually every sort of human expression—including blogposts, 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," wrote OpenAI in the House of Lords submission.

OpenAI claimed that the authors in that lawsuit "misconceive[d] the scope of copyright, failing to take into account the limitations and exceptions (including fair use) that properly leave room for innovations like the large language models now at the forefront of artificial intelligence."

you are viewing a single comment's thread
view the rest of the comments

[–] [email protected] 3 points 10 months ago

The amount of second hand content an LLM needs to consume to train inevitably includes copyrighted material. If they used this thread, the quotes OP included would end up in the training set.

The amount of fan forums and wikis on copy written material provide copious amounts of information about the stories and facilitate the retelling. They're right that it is impossible for a general purpose LLM.

My personal experience so far though has been that general purpose and multiple modality LLMs are less consistently useful to me than GPT4 was at launch. I think small, purpose built LLMs with trusted content providers have a better chance of success for most users, but we will see if anyone can make that work given the challenge of bringing users to the right one for the right task.