There's a problem with new technologies, in that cutting-edge technology develops quickly—in the span of years. Meanwhile, us human beings with our wet flesh bodies and our complicated bureaucracies and legal systems are glacially slow. You can see this playing out with generative AI.
For a few years now, we've heard a lot of overexcitement from the videogame industry that AI is going to let smaller companies make bigger games, that soon you'll be able to generate your own concept art, your own music, your own videogame dialogue—and, of course, the deeply tedious phrase "democratising art".
In actuality, generative AI is more of a plagiarism machine, often spitting out carbon copies and, even when it's not doing that, still requires countless copyrighted works (which sometimes have to literally be burned) to actually function. That legal heat has spooked videogame companies, so says lawyer Haley MacLean in an interview with our friends over at GamesRadar+.
MacLean, both a corporate IP lawyer and a videogame specialist at Voyer Law, tells the site that no-AI clauses are on the rise, particularly in the last year: "It's turned around, especially in the past year. I would say about two to three-ish years ago, you'd see a little bit of it … Even the smaller, more simple publishing studios are like, 'Let's include this just to cover our butts.'
"So now it's just boilerplate, to the point that if I did encounter one that didn't have a gen AI clause, if the publisher was creating assets too, I would say we need to push for that to make sure the publisher is not using gen AI in the marketing, porting, QA of your game."
It's not all too surprising, even from a ruthless capitalist perspective. There's a tension between videogame platforms, who are incentivised to rid themselves of the tides of AI slop lapping at their shores, and developers themselves. The former is encouraged to add large, obvious disclaimers so their users can filter out the garbage, while the latter is utterly screwed when they do.
Back in June, for instance, a data analyst found out that AI disclosures could reduce the number of reviews a game got by 53%, making the reviews more negative overall, which MacLean says causes a pervasive attitude of "I don't want to use gen AI because my community would hate it."
But the legal quagmire is also becoming increasingly insurmountable as legislation catches up—one particular quirk? If you use generative AI to make something, it might be impossible to prove that you copied the thousands of artists who were fed into your algorithm, sure. But it also means that, legally speaking, you can't really argue YOU made it, either.
"Sometimes I feel like I'm going crazy when I see these massive corporations … It kind of feels like a lawyer is hovering slightly over them, because they'll say things like, 'Look at this awesome tool. I can just type in 'barrel' and a barrel appears. But I have to make sure to change it a little bit!' They sort of imply that the users should be transforming every generated asset a bit, but they know that their users aren't going to do that."
MacLean uses the example of the Tomb Raider: Legacy of Atlantis remake, which uses generative AI for its assets—the game's not out yet, so it's unclear if it'll meet this bar, but if it doesn't change those assets enough? "[if there was] an outdoor level and all the trees are AI-generated, I legally would argue somebody could go and use every single tree in the Lara Croft game, that they know is AI-generated, and put in their game, and I don't think those devs can turn around and say, 'Those are our trees.'
"No, they're not, because you didn't generate them, and legally speaking, you have no ownership title over them if you didn't put in the sweat equity to have an artist go in and, inspired by that original first pass of the tree, then transform it into an original tree asset … But I think these major companies who are always laying off, trying to cut the bottom line, increase the value of the company, they're going to cut corners when they can, and I think they're going to get sloppy."
It's sort of a 'snake eats its own tail' situation. In order to fulfil the promise of generative AI—that you could lay off a bunch of artists and just have your hallucination machine dream something up—you need to lower your headcount. But your remaining artists then need to meaningfully transform what AI generates, in order to legally lay claim to it.
And if there's not only fewer of them, but they also kinda resent you for laying off all their colleagues, then that increases the chance of legal trouble happening—or lost sales from just pissing your consumer base off. At which point, you might not even be saving the money that a handful of yearly salaries (which are pennies to the absolute billions these companies have) saves you by way of their absence.
"I think," MacLean says, "We're going to see lawsuits of those types in the coming years when they want their protection but they want to do everything faster and cheaper … What case law in the US and Canada has seen, time and time again, is that AI cannot own because it was not human-made.
"We're in that frustrating middle with new tech where it's very fun, it's very flashy, everyone's excited, but I think the legal issues are going to just be popping up more and more."
