Nice Implementation of Something, but Not a Solution

[jcolag]

I've tried to wrap my head around what this wants to do, and coming up fairly empty.

Don't get me wrong, here. I also wish that we lived in a world where sprawling multinational corporations didn't see the Internet and its users as their property to monetize as they see fit. And I get it, Creative Commons in its heart of hearts, born in the fires of lawsuits threatening music sampling and Napster, while legislation and treaties threatened the public domain, doesn't like copyright or using it as a tool. I also understand that this means to sit downstream from a check that would ask AI scrapers to go away.

And yet, among the big companies, we've seen them argue that they can't respect copyright and survive, and by the way, art has no actual value, anyway. And I can remember not that long ago when a major player in this space insisted that this scraping constituted Fair Use, anyway, so the arguments didn't matter.

My first big problem, then, stems from how this creates a system where we pretend that, rather than sprawling corporations exploiting everybody, we actually have a both-sides etiquette problem. Moreover, the "solution" wants to put the burden on every creator ("starving artist," if you will) to ask politely for billion-dollar companies to maybe - if they have the opportunity and goodwill at some point - maybe consider that a website's maintainer doesn't want to help them externalize their costs.

After that, we have the redundancy of asking for attribution for a second time. But this time, rather than the legally binding, contractual requirement of attribution in exchange for use of the material, this version only requests a vague "best effort," with no legal backing to it whatsoever.

Minor, almost petty issue: If this would help at all, credit+1 seems incredibly silly. I realize that you might have a fear of proliferating edge-cases like the early modular Creative Commons licenses, but if you haven't tried to deny scraping upstream to Signals, it seems entirely sensible to ask for helping keep things sustainable and openness from the same organization (whereas no-derivatives and share-alike didn't make any sense together), especially when the requests get phrased to make them more aspirational than required.

On top of that, to put it bluntly, Creative Commons has already trashed the initiative's reputation. Especially with the "stop asking us, it's just Fair Use" aproach, as other people have pointed out, this comes off as saying that the future of Creative Commons lies in siding with billionaires who want to replace artists with software that lightly remixes existing art.

Likewise, a lot of people have already suffered the damage from these companies - and again, the companies cause the problems, not the specific technology that they use - so "did you consider just asking Meta to care about your work as more than a resource to exploit?" feels like victim-blaming. Maybe if this had come out two or three years ago, it'd seem more legitimate, but at this point, it doesn't look like more than a weak attempt to capitalize on anti-AI anger while not actually doing anything that will slow down the companies.

As an implementation standing alone, I don't see a problem with what it does. But it also actively refuses to actually accomplish anything - over and above the "doesn't actually do anything" aspect of robots.txt, because the 1990s didn't have companies stomping around openly saying that the rules shouldn't apply to them and going to court to defend that stance - and creates incredibly bad optics around the Creative Commons brand.

Having seen how these sorts of half-baked permission systems generally pan out, I'd say that, at best, these rules will only intimidate a couple of students into kicking in a couple of bucks to some organization claiming to represent "the ecosystem." But it'll also scare away a bunch of kids who can't afford to help and don't want to offend anybody. And it'll have absolutely no effect on the wealthy creeps who already see the entire Internet as an exploitable resource with no governance.

My recommendations:

• Stop worrying about pleasing everybody and find somebody, even one specific person, to care about in all this, and figure out how to solve their problem first.
• Focus on stopping the offenses, not teaching self-defense to the victims.
• Focus on stopping bad corporate behavior, not technologies, and certainly not buzzwords for technologies.
• Never assume that the people who keep repeating that the rules don't apply to them will follow the rules.
• Maybe accept that, when any AI chatbot can reproduce any license in full and other texts (it takes more prodding, these days, but try it), that the licenses actually have at least some relevance, because the language models might not have the perpetual Fair Use claim that you want.

I don't know what the correct solution looks like, but it probably doesn't look much like holding up a "Please Don't Mug Me" sign.

[CC-signals]

I don't know what the correct solution looks like, but it probably doesn't look much like holding up a "Please Don't Mug Me" sign.

We speak more to how we might incentivize compliance in this FAQ response.

First, there are contexts in which CC signals are likely to have legal effect. For example, the European Union’s legal regime on copyright and AI, particularly the Directive on Copyright in the Digital Single Market and AI Act foresee the possibility for rightsholders to opt out of the text-and-data mining (TDM) exception. In those cases, rightsholders reserve the right to conduct TDM and model developers need permission to use copyright works for AI training (see Article 4(3) of the EU Directive on Copyright in the Digital Single Market (EU 2019/790)). CC signals may serve as an express reservation, i.e. an opt out mechanism in this context, however there is still more to be determined.

Second, we have been exploring other measures to incentivize compliance, such as a petition to AI developers; an accreditation system where AI developers who follow CC signals get a special designation; and more.

[rebeccakross]

Coming back to this comment - first of all, thank you for writing it up. Your recommendations are fairly spot on. Specific use cases really do need to come next. And we are certainly hearing that no one is buying that AI companies are going to voluntarily comply. No easy answers here, but you've raised a number of issues we need to grapple with.