00:06
Adam Stofsky
If you're actually a creative, if you're an artist, maybe you're a composer. And briefly, we work with lots of artists, lots of animators, illustrators, story, we're artists, composers. And I'm sure some of these folks are thinking about using these tools. What do they need to think about when entering this world of generative AI and using these tools? 

00:29
Annette Hurst
Right. Well, a lot of artists are already familiar with using highly sophisticated digital art tools. Right. And what's perhaps next level about these generative AI tools is the degree to which they can ideate for artists as opposed to simply implementing. Right. They can come up with things that are conceptually appealing in ways that probably differentiate them from existing tools. But artists can and others can avoid creating greater risks by thinking about the inputs, the prompts in the lingo, prompts that they use to generate outputs. And one easy example is don't use somebody else's name in your prompt. Don't say in the style of so and so, because that's more likely than to result in something that could be problematic in the output. 

01:27
Adam Stofsky
But don't artists do that all the time? I'm not an artist. I just work with a lot of artists. But I might say to my artist, hey, can you draw me something in the style of this graphic novel, for example, or this some kind of pop culture reference? I mean, they do that all the time. Is there something different here, or is it still the same kind of caution you want to exercise? Do it in the style of don't copy it directly. 

01:54
Annette Hurst
Yeah, that's a great question. I mean, there's just a huge debate about that right now. Right. There are some artists who are, you know, like the technology, and they're excited about it and they want to use it. And they point out that the historical process of learning how to be an artist is from looking at what has come before you, learning from it, imitating it, and then changing those imitations enough that what you've done is actually create something new. Right. And that, and folks say that's exactly what AI models are doing. They're learning from the existing data in the same way that humans do and then helping you create something new. The other side of that argument is that it's completely different in terms of its nature. And the people who think it's different say usually two things. 

02:46
Annette Hurst
One, a machine software model does not have intent. It's not sentient. It's not learning the way humans are learning. And they argue that should make a difference. Other things that folks argue against it is that it's incorporating a massive amount of material that is far beyond what any individual human being could ever really use. And so it's unfair in some way to do that. And so it's also unfair for people to use that tool. And so there's a, you know, what I've just described as probably more of an ethical debate than a legal debate about, you know, amongst artists about this tool. 

03:29
Adam Stofsky
So from a legal standpoint, is there still a lot of uncertainty here? What, what the risks are for artists? 

03:38
Annette Hurst
I would say this. I would say for artists using the tools. You know, I think many artists are sophisticated enough to understand when they've gotten too close to some pre existing material versus created something new. So I think that artists who are attentive to those issues understand how to use tools properly, and that's less the issue, I think where the battle lines are drawn is in creating the tools in the first place, and that is really where the fight is to come.