Following up on being mentioned in a footnote in the document, amongst some 10K other commenters, I wanted to jot down initial notes and reactions as I skim through the rest of the document.
Page iii has key takeaways, and begins with recommendation of no change needed to current US copyright law. But the second point is, in my opinion, as clear as mud:
The use of AI tools to assist rather than stand in for human creativity does not affect the availability of copyright protection for the output.
Perhaps it will get clearer later in the document, but “assist rather than stand in for” is so… well, let’s just leave it at unclear for now. But this exact bridge is near where I’ve been trying to tie my horse (metaphorically) these past months in my work creating low tech human-powered robots and drawing machines without any electronics or motors or software. I think people who haven’t gotten out and explored this terrain carefully from an artistic/observer POV on the actual ground are much more likely to think there’s some clear either/or demarcation bright line between these two states, but there’s simply not. It’s a continuum with robotics and mechanical drawing machines, just like it is with generative AI… just how many linkages in a drawing arm do I need to add – precise count, please! – before the resulting artwork becomes no longer copyrightable? When does a tool become a stand-in? This is an extremely blurry – and I think vital! – question to explore more deeply, rather than say case closed…
This is another high-level takeaway, same page:
Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis.
I understand why, in their framework, this is likely the case. But to me as someone who has had to both write & enforce content policies, it sort of suggests that maybe they don’t have clear enough rules to apply. Often-times doing things on a “case by case basis” in a large queue workflow like they must certainly have for new copyright submissions, it can mean that you end up re-inventing the wheel again and again. Maybe that’s not the case here, but glossing over it as an outsider, that’s the impression I get. It doesn’t, at the very least, *sound* very efficient. [Insert obligatory Dept. of Govt. Efficiency reference… btw here is its titular head’s take on AI & copyright… make of it what you will.]
On that note, I’ve grown more accustomed and see the wisdom in Canada’s arrangement where the IPO does not review or assess submissions, everything is automatically approved. If it’s contested, it is a matter for the courts. Given that so many copyright cases go to court anyway, in one way of looking at things, it seems possibly more efficient to just let it be decided there on a case-by-case basis, rather than having this other secondary body doing *another* layer of case-by-case analysis, which itself becomes open to being challenged in court.
Well, seeing as how wound-up I got only just off a few lines here, this might be a wise place to stop for the night.
Leave a Reply
You must be logged in to post a comment.