Copyright. The right to copy?

There’s been a lot of discussion on the interwebs the last couple of days about copyright and AI, much of it centred around this article, where OpenAI claims they cannot create an innovative tool like ChatGPT without access to copyrighted material (cue sobs and tears).

On the one hand, OpenAI is a well-funded, Silicon Valley company, and I, for one, have so little sympathy for their claim that one would need a pretty powerful electron microscope to see it. They have no qualms about exploiting the creations and labour of others in order to entice you and me, and well-moneyed entities like Microsoft to pay handsomely for the privilege of using the tools they have developed.

My gut feeling is that if they need to use copyrighted material, they can bloody well pay for it. Like everyone else does. And if that means they can’t afford to use as much source material as they would like, that’s just too bad. There are many tiny violins ready to play for them.

But then, on the other hand, I think back to a time around 20 years ago when Napster, Limewire, and the Pirate Bay were the enemies du jour of the copyright industry — in that case, the major record labels who owned and controlled access to music pretty much worldwide. I cheered for the pirates, not because I wanted to destroy the possibilities artists had to make a living off their art, but because the gate keepers were the greater evil, stopping people from accessing music unless they went through the gate keepers.

And I am conflicted today. OpenAI is not the Pirate Bay, a small group of idealistic hackers wanting to undermine a monopolistic, capitalist gatekeeping system that locked artists behind paywalls, but are there similarities? Given the funders behind OpenAI, I am inclined to think not.

But I still need to think about it.

AI and ethical conundrums

I am one of a group of faculty members who have organised a symposium on «AI and Art» at our university on January 12th. It seemed natural for me; I have always been interested in the ways artists explore new technologies, and push them in ways that perhaps was not intended. This was why I led an initiative that leaned heavily into Virtual and Extended Reality, despite some significant scepsisism from many of the filmmakers who teach at the Norwegian Film School.

New technologies always come with baggage; I don’t believe there is such a thing as a completely neutral technology. There is always a worldview, as set of assumptions and values, and both a cost and a benefit to any technology / tool.

But rarely has this been so evident as in the case of the AI tools that have dominated public discourse in the past 18 or so months. These are tools that can do wonderful things (no link; you can find them easily enough if you haven’t seen them already) but they also come with a significant price attached. We’ve seen the reports of the Kenyan content moderators, exposed to the worst the web has to offer for what amounts to slave wages. We’ve seen the examples of racist and sexist biases build in to the tools.

The biases and prejudices built into the tools are one thing. The other side is it became apparent early on that the companies that have created the AI tools have asserted a right of fair use to copyrighted creations on the web, and copyright holders — including writers, artists, musicians, and more — have launched lawsuits asserting their intellectual property (oh, how I loathe that term) has been stolen.

This means every time, or another artist, uses a generative AI tool like ChatGPT or Midjourney, we are, in fact, using works we do not have to right to in order to create our own.

But is it that simple?

I neither know nor understand all the legal implications of this, but I do know that in the art form I work in — filmmaking — artists have been «stealing» from artists they admire almost for as long as there has been a film industry. Would we have a Kurosawa ouevre to admire if he had not stolen from John Ford? Would we have a Star Wars franchise (that never seems to end) if George Lucas had not stolen from Kurosawa (and many others)?

I know there is a difference between artists taking inspiration from others they admire and a corporation almost mindlessly harvesting artistic works in order to make their machine as slick as possible. But it still gives me pause for thought.

2 of 366?

I don’t make New Year’s resolutions, but am motivated to get into a writing habit. Yesterday, I was part of a discussion about it on Mastodon, started by Kate Bowles where I jumped in and suddenly realized — inspired by yesterday’s post — that I want to write. I want to be one of the people who find writing a necessity and something one must do in order to think properly. We’ll see if it actually becomes a daily habit, but for now, I will try.

But why?

Much of it has to do with my field — pedagogy for fine arts filmmaking — and some very interesting things I’ve been reading lately about “4E cognition”¹ and especially how it relates to creativity. Also, I’ve been reading a lot about the various so-called AI tools that have appeared in the public eye over the past 18 or so months.

The confluence of these things is an interesting thing to muse about. It also necessitates, I think, a concerted effort to, well, think. Writing, as many people have pointed out, most recently (for me), Sarah Honeychurch, writing is thinking and is the one thing we cannot — must not — outsource to a machine.

Thinking, creativity; these are part of what makes us human, what makes us conscious. And while AI
can help us in many ways², and become a tool that enables forms of creation we perhaps did not have access to previously, it cannot replace human creativity (at least, not yet!).

Footnotes

In Schiavio, Andrea, and Dylan Van der Schyff. 2018. “4E Music Pedagogy and the Principles of Self-Organization” Behavioral Sciences 8, no. 8: 72. https://doi.org/10.3390/bs8080072, the
authors define 4E as: – Embodied: Cognition cannot be fully described in terms of abstract mental processes (i.e., in terms of representations). Rather, it must involve the entire body of the living system (brain and body). / – Embedded: Cognition is not an isolated event separated from the agent’s ecological niche. Instead, it displays layers of co-determination with physical, social, and
cultural aspects of the world. / – Extended: Cognition is often offloaded into biological beings and non-biological devices to serve a variety of functions that would be impossible (or too difficult) to be achieved by only relying on the agent’s own mental processes. / – Enactive: Cognition is conceived of as the set of meaningful relationships determined by an adaptive two-way exchange between the biological and phenomenological complexity of living creatures and the environments they inhabit and actively shape.↩︎
Assuming the companies that have created the most high profile AIs survive the barrage
of lawsuits being launched against them by the artists whose work they have stolen…↩︎