265
OpenAI says it’s “impossible” to create useful AI models without copyrighted material
(arstechnica.com)
this post was submitted on 11 Jan 2024
265 points (100.0% liked)
Technology
37720 readers
32 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
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
If it is impossible, either shut down operations or find a way to pay for it.
My concern is they and other tech companies absolutely can and would pay if they have no choice. Paying fines for illegal practices if needs be.
What absolutely wont survive a strong law to keep copyright content out of ai is the open source community which absolutely can not pay for such a thing and would be seriously lacking behind if its excluded, Strengthen the monopoly on ai by for Profit Tech. So basically this issue can have huge ramifications no matter what we end up doing.
My understanding of the open source community is that taking copyrighted content from people who haven't willingly signed onto the project would kind of undermine the principles of the movement. I would never feel comfortable using open source software if I had knowledge that part or all of it came from people who hadn't actively chosen to contribute to it.
I have seen a couple of things recently about AI models that were trained exclusively on public domain and creative commons content which apparently are producing viable content, though. The open source community could definitely use a model like that, and develop it further with more content that was ethically obtained. In the long run, there may be artists that willingly contribute to it, especially those who use open source software themselves (eg GIMP, Blender, etc). Paying it forward, kind of thing.
The problem right now is that artists have no reason to be generous with an open source alternative to AIs, when their rights have already been stomped on and certain people in the open source community are basically saying "if we can't steal from artists too, then we can't compete with the corporations." So there's literally a trust issue between the creative and tech industries that would need to be resolved before any artists would consider offering art to an open source AI.
Its quite a mess but I definitely agree that open source needs a good model trained on consented works.
I do fear though that the quality gap between copyright trained and purist models will be huge in the first decenia. And no matter the law, the tech is out there and corporation and criminals will be using it in secret nonetheless.
If only things where as simple as choosing for the chad digital artists. Digital art was part of my higher education and if i Haden t get a tech job i might have been one of them so i feel torn between the divide in industries.
This may sound doomer but since the technology exist we are in a race to obtain beyond human super intelligence and we do not know what will happen after that.
OpenAI had multiple times stated they don’t know if copyright will still mean anything in a future with ai.
We are also facing some huge global issues like global warming where a super intelligence could be the answer to sustain the planet, of course also risking evil ai in the process… i repeat such a mess
I don’t fully trust sam altman, but i do believe what they say may be true. At some point its going to be here and it will be to smart to ignore.
Its optimistically possible that in 20 years we will all be leisurely artist laughing at the idea of needing to work to earn survival.
Its of course just as likely some statehead old bastard presses the deathbutton next week and thats the end of all of it or that climate has progressed beyond what our smartest future ai could possible solve.
I definitely do not have the optimism that in 20 years time we'll all be leisurely artists. That would require that the tech bros who create the AIs that displace humans are then sufficiently taxed to pay UBI for all the humans that no longer have jobs - and I don't see that happening as long as they're able to convince governments not to tax, regulate, or control them, because doing so will make it impossible for them to save the planet from climate change, even as their servers burn through more electricity (and thus resources) than entire countries. Tech bros aren't going to save us, and the only reason they claim they will is so they never face any consequences of their behaviour. I don't trust Sam Altman, or any of his ilk, any further than I can throw them.
That's is why i am putting some of my eggs in open source, which is where the real innovation happens anyway. Free Ai tools at home running on consumers devices can level people up to build a better future ourselves without having to rely on techbros or government.
Of course i should nuance my wording a bit. My actual opinions tend to be contrasting mix of both optimistic and pessimistic lines of evens. I dont have much hope that the good future is the one we will end on, but it remains in my speculative opinion possible from where we are standing today, yet all can change in less than a week.