We’re now into the third week! 8 out of 16: we’re halfway through. We sincerely hope you’ve been enjoying the reading so far.
Today’s keyword is “legality”. How can one navigate copyright laws, especially when they vary from one geographical region to another*? The issue of copyright can quickly become a headache and discourage the sharing of resources for fear of making a mistake that could prove costly.
Rory McGreal (@Rory) approaches this issue from a particular angle: copyright as it applies to the use of generative AI. @Rory aims to “demystif[y] the legal landscape”, offering insights designed to facilitate the use of generative AI in the creation of OER, whilst highlighting that “a clear legal trend is emerging that strongly supports openness in education”.
We let you find out why in more detail in this article.
(* This article turned out to be particularly difficult to translate for this reason.)
Who Owns AI-Generated Content? by Rory McGreal
Perhaps the most famous sharer of all time is Robin Hood. But he was also an outlaw. So it’s quite normal that we should all be concerned about sharing today. What is the reality? When do I have the right to share? And above all… what changes with AI? These are the questions addressed by Rory McGreal (Athabasca University, Canada) in today’s topic.
Much appreciation sending to @Rory for encapsulating so clearly the certain and uncertain of GenAI legality. It’s profound how much and how quickly it has shifted much into the zone of questioning.
With not any kind of deep research and purely a personal response…
What does it really mean to own an idea or an expression of it? This alone has been a block on the idea of sharing we form as children and leads us to the Gollum inspired columns previously. I hold firm to the school of Kirby Ferguson that everything is a remix. While the “how” of how GenAI creates by drawing connections from a wide range of information is different, at the same time we as humans are always doing something like that. What if we never held to ownership? People want to be credited for their works, is the only means to do that a stamp of ownership?
The ides of a copy is outdated. Aren’t we operating from the more familiar idea of copy as a duplication, the veritable digital xerox machine? I feel we are hampered to understand what GenAI does by thinking it is working by copying.
The paradox of and over focus on licenses Open licenses especially CC has enabled so much especially to bust open the copyright bind, but at the same time I feel we go on endless about quibbles of licenses, and the tendancy to suggest to creators that a license is some kind of protection.
The Grand Washing Machine One might say that the vast majority of content created and placed out into the world, copyrighted or night, is all consumed and returned to us now as public domain? Maybe not a bad thing, yet at the same time there has been numerous examples where LLMs have ben shown to reproduce close approximating of original content from all it has ingested. But if we were not so tied to owning, would it be a problem? And what happens beyond the false idea of GenAI vs Human, when it is all mixed and remixed with itself?
What a great series this has been @erwan_louerat and colleagues, and we are but partly over. I have been keenly following a few of the efforts to make the transparency clear and am looking forward to an OEWeek Session tomorrow by @agrey at Kwantlen Polytech on their approach to an AI Declaration Statement for OER.
And speaking of sharing, I would give up ownership of everything if more people would enter these discussions.