Can You Copyright AI-Generated Work? | AI Lawyer Lisa Oratz Explains

WEBVTT
1
00:00:00.672 --> 00:00:07.317
Hi, and welcome to Technically Creative, where we talk to creative leaders about how technology and AI are changing their industries and their companies.
2
00:00:07.917 --> 00:00:15.323
This episode is brought to you by the Future of Life Institute, an independent nonprofit working to steer transformative technologies towards benefiting humanity.
3
00:00:15.703 --> 00:00:19.746
Senior Council of Entertainment, IP, and Technology at Ashurst Perkins Coy.
4
00:00:20.367 --> 00:00:29.254
And I think this is one of the most practically useful conversations we've had for anyone making creative work with AI, because there's an enormous amount of confusion about copyright right now.
5
00:00:29.814 --> 00:00:31.717
Can you copyright something made with AI?
6
00:00:32.238 --> 00:00:35.943
Does AI—does using AI mean you lose ownership of your work?
7
00:00:36.384 --> 00:00:40.410
If you generate a character, do you own the character, or do you only own the film you've put it in?
8
00:00:40.830 --> 00:00:44.396
And how much human involvement is enough to make something legally protectable?
9
00:00:44.796 --> 00:00:46.697
These aren't hypothetical questions anymore.
10
00:00:46.777 --> 00:00:52.301
They're questions that filmmakers, advertisers, studios, and independent creators are having to answer every day now.
11
00:00:52.482 --> 00:00:54.303
And Lisa's been helping people answer them.
12
00:00:54.803 --> 00:01:03.469
She worked on Critters, the first AI-generated animated short film to receive copyright registration that then put it on the path to becoming a feature-length film.
13
00:01:03.710 --> 00:01:12.356
And what I find fascinating about that case is that it challenges one of the most common assumptions about AI-generated work, that if you made it an AI, you can't own it.
14
00:01:12.976 --> 00:01:15.480
It turns out the answer is more interesting than that.
15
00:01:15.961 --> 00:01:20.908
Because copyright isn't about which tools you used, it's about human authorship, as it ever was.
16
00:01:21.509 --> 00:01:27.358
The decisions you made, the elements you created, what you selected, what you changed, and how you brought everything together.
17
00:01:28.019 --> 00:01:36.383
And while the law is still developing, Lisa makes a compelling case that professional creatives may already be better positioned to protect their AI-assisted work than they realize.
18
00:01:36.723 --> 00:01:38.424
But there's another side to the conversation.
19
00:01:38.784 --> 00:01:47.027
Because being able to protect something you've made with AI is very different from knowing whether the AI companies were entitled to use the material they trained their models on in the first place.
20
00:01:47.427 --> 00:01:57.592
And this is where we get into the well-worn conversation of the fair use argument, the anthropic and meta cases, and why two courts looking at very similar questions have reached some surprisingly different conclusions.
21
00:01:57.972 --> 00:02:04.777
I think one of the most important things Lisa does in this conversation is separate what we actually know from what people confidently claim to know.
22
00:02:04.797 --> 00:02:06.558
Because we're in a strange position.
23
00:02:06.878 --> 00:02:09.140
The technology is already being used professionally.
24
00:02:09.440 --> 00:02:11.121
The creative work is already being made.
25
00:02:11.482 --> 00:02:13.383
The business models are already changing.
26
00:02:13.803 --> 00:02:16.886
But some of the most fundamental legal questions remain unresolved.
27
00:02:17.506 --> 00:02:20.228
So what are creative people supposed to do in the meantime?
28
00:02:21.058 --> 00:02:33.248
We get into the practical steps creators can take to protect their work, why documenting your creative process matters, how to reduce the risk of infringement, and what the emergence of synthetic performers and digital replicas might mean for the entertainment industry.
29
00:02:33.428 --> 00:02:37.752
Because I don't think the future of creativity is going to be decided by whether we use AI or not.
30
00:02:38.292 --> 00:02:45.398
It's going to be shaped by how we use it, what we choose to make, and whether the people doing the creative work can retain meaningful ownership over what they create.
31
00:02:45.818 --> 00:02:48.961
And Lisa brings some much-needed clarity to exactly those questions.
32
00:02:49.461 --> 00:02:50.362
This is Lisa Oratz.
33
00:02:59.128 --> 00:03:01.710
Lisa Oratz, thank you so much for joining us on the podcast.
34
00:03:01.750 --> 00:03:03.252
We've been trying to do this for a while.
35
00:03:03.532 --> 00:03:04.673
Thanks for being on the podcast.
36
00:03:04.913 --> 00:03:05.754
Well, thanks for having me.
37
00:03:05.774 --> 00:03:07.155
I'm very excited to be here.
38
00:03:07.256 --> 00:03:11.860
Well, Lisa Oratz, you are senior counsel at Ashurst Perkins Coie.
39
00:03:12.100 --> 00:03:19.967
And you first came on my radar when I saw you speak at AI on the Lot, which is a great conference out in LA about technology in entertainment.
40
00:03:20.067 --> 00:03:24.832
We have viewers across film and television and advertising for someone outside the legal world.
41
00:03:25.032 --> 00:03:30.936
What did they need to understand about what you do and who you advise and the problems that you're helping clients solve right now?
42
00:03:31.116 --> 00:03:36.279
So my work covers three areas, technology law, entertainment, and intellectual property.
43
00:03:36.699 --> 00:03:40.942
So my clients range from big tech companies to little startups.
44
00:03:41.282 --> 00:03:47.426
But I also spend a lot of time working with studios, production companies, and artists on the entertainment side.
45
00:03:47.826 --> 00:03:50.828
And my real special sort of fun spot is where these all collide.
46
00:03:51.328 --> 00:03:56.271
And some of the issues that we were talking about AI on the lot is really at that intersection.
47
00:03:56.371 --> 00:04:04.756
So I advise a lot of clients on how to address some of the issues and navigate this uncertain legal landscape on IP issues.
48
00:04:04.796 --> 00:04:14.041
But I also work on deals, contracts, you know, really a wide range of work for, you know, entertainment, advertising, gaming and tech clients.
49
00:04:14.661 --> 00:04:20.825
What are you seeing most people get wrong when they only have a surface level understanding of the copyright challenges that we're facing here?
50
00:04:20.989 --> 00:04:26.152
Yeah, and there is a lot of misunderstandings and misinformation out there.
51
00:04:26.192 --> 00:04:30.015
It's a very complicated area, and everybody jumps in and wants to be an expert.
52
00:04:30.275 --> 00:04:40.541
But there's two key things that I really want, and we can dive down in more detail as we talk, but really want people to come away with is, you know, there's the copyrightability and protectability aspect.
53
00:04:40.561 --> 00:04:50.107
And there were a lot of articles and fear out there with respect to some of the recent decisions came out that says, you know, AI-generated work isn't protectable.
54
00:04:50.567 --> 00:04:52.348
But that really is a misunderstanding.
55
00:04:52.388 --> 00:04:55.170
There's a lot of question about exactly what's protectable.
56
00:04:55.670 --> 00:05:01.713
And in particular, whether the actual output that comes out of an AI tool is protectable.
57
00:05:01.774 --> 00:05:03.495
But the law is still developing on that.
58
00:05:03.955 --> 00:05:09.038
And even with that uncertainty, and frankly, a copyright office that is a little bit hostile to AI,
59
00:05:09.658 --> 00:05:11.279
The general works as a whole.
60
00:05:11.299 --> 00:05:18.223
So if you're whatever you're creating, as long as there is sufficient human authorship, which is the big question, these things can be protectable.
61
00:05:18.323 --> 00:05:19.844
So that's one misunderstanding.
62
00:05:20.364 --> 00:05:21.905
And the other one is around infringement.
63
00:05:21.925 --> 00:05:24.606
I mean, there's over 100 cases that have been brought now.
64
00:05:25.047 --> 00:05:33.071
But from a user standpoint in particular, there's a lot of things that users can do to really avoid and minimize any risks of infringement.
65
00:05:33.611 --> 00:05:36.153
And again, it's an area where with a lot of uncertainty.
66
00:05:36.173 --> 00:05:36.473
Right.
67
00:05:36.693 --> 00:05:45.196
But where you see a lot of alarming headlines, particularly with some of the early head, there was a recent settlement with Anthropic, which we can talk about.
68
00:05:45.576 --> 00:05:51.058
But there's a lot of misunderstanding exactly what the parameters of that decision means.
69
00:05:51.078 --> 00:05:54.499
So when I talk at these conferences and podcasts and things like this.
70
00:05:55.079 --> 00:06:05.964
My goal is to really help people understand the line between what we know about the law, what we don't know, and how you can navigate that in an environment that, frankly, is going to be uncertain for quite some time, likely.
71
00:06:06.264 --> 00:06:17.309
I think that there's obviously the fundamental copyright provenance issues with the models themselves, their training data on what's called fair use, the fair use argument that...
72
00:06:17.749 --> 00:06:25.839
similar to if I were to read a book and then be inspired by that book and then create a book myself, that's a fair use of that copyright.
73
00:06:25.899 --> 00:06:36.632
And then the secondary point that you're making as well, which is then what happens when in a human workflow, that stuff that comes out,
74
00:06:37.433 --> 00:06:39.774
You mentioned the hostile copyright office.
75
00:06:39.934 --> 00:06:48.637
What does one need to do to actually have ownership and have an IP protection or copyright protection around that outcome of that work?
76
00:06:49.077 --> 00:06:52.718
So maybe taking those sort of in reverse order, perhaps.
77
00:06:52.998 --> 00:06:59.920
So one of the key issues and probably the most concern that I hear from people is if I use AI, am I going to lose protection?
78
00:06:59.940 --> 00:07:04.402
Am I going to be able to protect and prevent other people from using my work?
79
00:07:05.082 --> 00:07:05.602
Excuse me.
80
00:07:05.982 --> 00:07:11.183
And the big issue here is that there does have to be some human authorship in order to have a copyright.
81
00:07:11.644 --> 00:07:15.665
And we're still trying to figure out exactly what is sufficient.
82
00:07:15.825 --> 00:07:18.245
If you have a very detailed prompt, is that enough?
83
00:07:18.685 --> 00:07:25.787
And that's a big question mark where the Copyright Office has said that a text prompt by itself is not sufficient.
84
00:07:35.209 --> 00:07:37.452
what gets registered, and they can be influential.
85
00:07:38.092 --> 00:07:47.141
But we don't know whether, based on a simple text prompt or even a complex text prompt, the actual content that spits out of the tool will be protected.
86
00:07:47.741 --> 00:07:54.048
But most of my clients, at least, don't just put in one prompt and out spits an ad or a film or a creative work.
87
00:07:54.648 --> 00:07:56.351
They are doing a lot more with that.
88
00:07:56.391 --> 00:07:57.272
They are taking that.
89
00:07:57.292 --> 00:07:59.435
They're combining different output.
90
00:07:59.815 --> 00:08:00.837
They're modifying it.
91
00:08:01.338 --> 00:08:10.050
And even with a pretty narrow view of protectability under the Copyright Office view, they have recognized that these works as a whole are protectable.
92
00:08:10.310 --> 00:08:18.598
And for example, we have worked on the film Critters and obtained copyright protection for that film, which was an AI generated short film.
93
00:08:19.038 --> 00:08:21.080
It's now being made into a full length feature.
94
00:08:21.100 --> 00:08:26.105
But we think this was the first actual copyright registration for an animated short film.
95
00:08:26.885 --> 00:08:29.788
And we were able to get protection on the film as a whole.
96
00:08:30.409 --> 00:08:36.174
We had the question mark again was whether, you know, which elements within that film were protectable.
97
00:08:36.581 --> 00:08:43.971
And you do have to describe when you file a copyright what the human authorship of that consists of.
98
00:08:43.991 --> 00:08:45.994
You don't have to go into great detail.
99
00:08:46.034 --> 00:08:46.755
So you don't have to.
100
00:08:47.350 --> 00:08:51.192
give every element of human involvement.
101
00:08:51.612 --> 00:08:58.997
But you do want to make sure that when you create works, that you document, that you insert a lot of human involvement.
102
00:08:59.017 --> 00:09:02.359
You want to make sure that you're taking work, you're combining that.
103
00:09:02.399 --> 00:09:16.687
Because the copyright office recognizes, and copyright law, I shouldn't say, recognizes the concept that even if you take unprotectable elements, when you make selections and arrangement of those elements and put it into an end product, that end product is protectable.
104
00:09:17.347 --> 00:09:22.750
So on that basis, it's not hard to get copyright protection for, you know, that output.
105
00:09:22.770 --> 00:09:24.411
And we've done this now several times.
106
00:09:24.871 --> 00:09:29.253
That's what I really want people to come away with is you can protect the output.
107
00:09:29.533 --> 00:09:37.277
And what we suggest in terms of trying to address the issue, well, what about the question of the output that comes directly out of a machine?
108
00:09:37.317 --> 00:09:39.038
Because theoretically, the concern is...
109
00:09:39.878 --> 00:09:44.859
If somebody were to extract that output and claim it's not protectable, could they use it?
110
00:09:45.139 --> 00:09:49.840
But you don't have to tell them what was the original output and what did you change and modify.
111
00:09:49.880 --> 00:09:55.341
So if you're taking output, you're combining it, you're changing it, you're putting it into your film.
112
00:09:55.741 --> 00:10:03.243
It's very difficult as a practical matter for somebody to figure out what might be unprotectable elements and be able to take them and use them.
113
00:10:03.303 --> 00:10:07.544
So I think a lot of those concerns as a practical matter really aren't a big deal.
114
00:10:07.944 --> 00:10:18.033
Yeah, I think that that's what I'm intrigued by your work is that most of the public conversation around AI and copyright is framed as AI work can't be copyrighted or it's tainted somehow.
115
00:10:18.073 --> 00:10:24.038
But your work seems to sit in the more important middle when AI is part of a human creative process.
116
00:10:24.439 --> 00:10:27.682
What evidence proves that the human contribution still matters?
117
00:10:27.822 --> 00:10:29.924
I don't know if our audience will remember, but Critter's.
118
00:10:30.584 --> 00:10:36.966
was certainly the first online piece of work that I had seen that now everybody would call slop.
119
00:10:37.606 --> 00:10:42.127
But what I take that to mean is that it was end-to-end AI creation.
120
00:10:42.347 --> 00:10:45.588
It was these cute little critters in a forest.
121
00:10:45.788 --> 00:10:50.169
And then it was the first thing to be optioned as an AI piece of work.
122
00:10:50.389 --> 00:10:53.270
And you created the copyright protections around it.
123
00:10:53.430 --> 00:11:11.250
I guess the question, and just using critters as an example, what was the human creative contribution or the meaningful human intervention that differentiated critters and made it both not slop, but also a professional project that could then be protected?
124
00:11:11.270 --> 00:11:11.450
Yeah.
125
00:11:11.851 --> 00:11:15.314
Yeah, because I definitely wanted to make sure we disagree with the characterization.
126
00:11:15.334 --> 00:11:18.137
This is a very high quality output.
127
00:11:18.157 --> 00:11:21.761
But the thing that was important is there was a lot of human input in that.
128
00:11:22.762 --> 00:11:26.005
And what's interesting about the short is we didn't have Zora.
129
00:11:26.045 --> 00:11:31.530
We didn't really have, what I should say, there weren't AI video tools that were broadly available.
130
00:11:31.550 --> 00:11:35.735
So that was actually created using DALI and using still images.
131
00:11:36.315 --> 00:11:42.117
So the images, the character images, the background images, those were all created through Dali.
132
00:11:42.717 --> 00:11:45.618
But once they had those images, they combined them.
133
00:11:45.738 --> 00:11:47.879
So, for example, there were a lot of background panels.
134
00:11:47.939 --> 00:11:52.280
They were put together, stitched together, modified, sprinkled with pixie dust.
135
00:11:52.380 --> 00:11:53.380
I love that.
136
00:11:53.420 --> 00:11:56.601
Literally, you could see that on there, and that was something we used to describe this.
137
00:11:57.061 --> 00:12:01.143
So there was a human that went through and did all of that work.
138
00:12:01.643 --> 00:12:04.844
And then the actual character images, those were all modified.
139
00:12:05.664 --> 00:12:08.666
using tools that a human was involved in.
140
00:12:08.706 --> 00:12:11.848
There was also human script, human actors.
141
00:12:12.089 --> 00:12:15.091
They captured, you know, so they did some voice capture.
142
00:12:15.111 --> 00:12:22.716
So there was an actual lot of human involvement in that, which is one of the things that made it fairly easy to obtain protection.
143
00:12:23.236 --> 00:12:27.918
What had to be documented then on their side in order for that to hold water?
144
00:12:28.118 --> 00:12:30.419
It's always good to internally document all that.
145
00:12:30.459 --> 00:12:39.083
But when you actually apply, a lot of people think you have to give a great amount of details to the Copyright Office is exactly what all that involvement was.
146
00:12:39.643 --> 00:12:42.284
And they really, the Copyright Office doesn't want a lot of detail.
147
00:12:42.304 --> 00:12:43.464
They want to know that.
148
00:12:43.484 --> 00:12:45.785
So we described it really just a paragraph or two.
149
00:12:46.486 --> 00:12:49.187
And sort of exactly what I just said was we explained how
150
00:12:49.847 --> 00:12:51.028
The images were generated.
151
00:12:51.088 --> 00:12:52.109
They were modified.
152
00:12:52.129 --> 00:12:55.551
They were selected and arranged in a particular manner.
153
00:12:55.591 --> 00:12:56.792
There was a human script.
154
00:12:57.633 --> 00:12:59.314
So we were able to describe that.
155
00:12:59.794 --> 00:13:05.138
And we always recommend, again, keeping as detailed records as you can in terms of all those human touchlines.
156
00:13:05.178 --> 00:13:06.119
Just because there's so much.
157
00:13:06.139 --> 00:13:07.660
Because we don't ask that.
158
00:13:07.740 --> 00:13:09.321
Is it a huge logging task?
159
00:13:09.381 --> 00:13:10.302
Is that what you mean?
160
00:13:10.762 --> 00:13:11.563
It can be.
161
00:13:11.583 --> 00:13:17.107
You know, because, you know, for many reasons, it's good to log prompts and to keep a record of that.
162
00:13:17.647 --> 00:13:23.774
But as you remember, the Copyright Office, as I said, doesn't really recognize the copyrightable, the output based on prompts.
163
00:13:24.394 --> 00:13:32.263
So the thing that's the most important is what you do with the output, how you modify it, how you combine it, and all of those.
164
00:13:32.283 --> 00:13:34.545
But I do think that all of it should be documented.
165
00:13:34.685 --> 00:13:38.129
We don't know exactly what's going to be the most important.
166
00:13:38.149 --> 00:13:38.249
Yeah.
167
00:13:38.369 --> 00:13:46.354
And one thing we did that was just interesting in Critters, and we've done this in subsequent applications as well, is you have to disclaim what's not protectable.
168
00:13:46.474 --> 00:13:49.235
The challenge is we don't know legally what's not protectable.
169
00:13:49.755 --> 00:13:57.520
But we know that if we didn't mention the actual output, if we tried to claim that, the Copyright Office would have rejected our application.
170
00:13:57.540 --> 00:14:02.563
But at the same time, we didn't want to exclude the AI-generated images directly.
171
00:14:03.263 --> 00:14:07.707
from that protection because, again, they could be wrong and we don't know until we have some case law.
172
00:14:07.727 --> 00:14:14.293
So what we did is we excluded AI-generated images to the extent not copyrightable.
173
00:14:14.453 --> 00:14:19.738
So it's kind of a circular argument to give it a little bit of room to flex as the law developed.
174
00:14:19.778 --> 00:14:22.020
We didn't really know what they did with it, but they accepted that.
175
00:14:22.100 --> 00:14:28.025
So, you know, in this case, at least, it's not, you know, there's going to be uncertainty about those issues.
176
00:14:28.365 --> 00:14:29.885
initial images that came out.
177
00:14:29.905 --> 00:14:34.046
So those are the initial character images, the initial background panels.
178
00:14:34.526 --> 00:14:44.888
But to the extent that we actually, and we should have a decision coming up sometime soon, because there was a recent copyright office ruling that had denied protection based on text prompts that was challenged.
179
00:14:45.528 --> 00:14:47.909
And so we may actually get a little bit of law on this.
180
00:14:48.449 --> 00:14:54.990
But until that happens, you know, we have at least a claim that, you know, even those text prompts are protectable.
181
00:14:55.290 --> 00:15:06.918
But the most important thing that I really want people to come away with is that as long as you make sure that you do include humans, you know, they talk about humans in the loop, but they're making decisions.
182
00:15:06.998 --> 00:15:13.242
They're shaping their because of what the Copyright Office view is that they want to make sure there's a human sort of controlling the output.
183
00:15:13.262 --> 00:15:16.024
So you want to show that there's a human being sort of in charge.
184
00:15:16.224 --> 00:15:19.765
Whether or not it has to reach high or threshold, we don't know.
185
00:15:20.125 --> 00:15:22.026
But if it does, that you show that.
186
00:15:22.046 --> 00:15:24.847
So we always suggest try to make sure you reach that bar.
187
00:15:24.867 --> 00:15:27.928
Hopefully we'll get some clarity that that's not needed.
188
00:15:27.968 --> 00:15:31.229
But the documentation, again, it's important, but you don't want to share that.
189
00:15:31.249 --> 00:15:40.912
You don't want to make it publicly available because you don't want people to know what came out of an AI tool and may or may not be protectable and what you add to it.
190
00:15:41.112 --> 00:15:45.815
Yeah, it's interesting because we have a lot of people who watch this who are creative professionals.
191
00:15:45.835 --> 00:15:57.741
So I think many of them will be encouraged by what you just said because they're used to working in very human-centric ways and producing high-level professional work that has lots of human intervention.
192
00:15:58.161 --> 00:15:59.602
And so I think that they're used to that.
193
00:15:59.642 --> 00:16:01.383
I think they'll be very comforted by that.
194
00:16:01.743 --> 00:16:03.464
They're used to having a business affairs department.
195
00:16:03.684 --> 00:16:05.866
that you're running these sorts of questions by.
196
00:16:06.246 --> 00:16:13.533
Whereas it's actually the new creators who are for the first time ever in their lives, having access to really highly developed post-production tools.
197
00:16:14.034 --> 00:16:21.421
They need to hear a lot of this as well to help them understand the rigor that goes into creating something that's professional and defensible.
198
00:16:21.656 --> 00:16:21.916
Yeah.
199
00:16:22.056 --> 00:16:30.241
And the other thing I want to mention that helps in terms of, again, maximizing your arguments for protection is start with stuff other than text prompts.
200
00:16:30.621 --> 00:16:36.185
And again, text prompts may be fine, but a lot of people start with sketches, storyboards.
201
00:16:36.685 --> 00:16:41.868
And to the extent that you start with something like that, that those are clearly copyrighted.
202
00:16:41.888 --> 00:16:42.929
If you create a sketch...
203
00:16:43.549 --> 00:16:44.550
You create a storyboard.
204
00:16:44.570 --> 00:16:47.233
That's a copyrightable document in and of itself.
205
00:16:47.673 --> 00:16:53.619
So if you're starting with that, to the extent that that's reflected in your output, that is going to be protectable.
206
00:16:53.759 --> 00:17:05.872
So that's an additional thing that people can do to make sure that they're maximizing exactly what's protectable is to start out just with text prompts, but with creative, what they call expressive prompts.
207
00:17:06.293 --> 00:17:09.697
And so that's very interesting because where do we then draw that line?
208
00:17:09.897 --> 00:17:12.700
Obviously, if I've drawn something or I'm a storyboard artist, great.
209
00:17:13.261 --> 00:17:16.264
But many production designers begin with a mood board.
210
00:17:16.744 --> 00:17:23.231
So if I bring to the LLM a mood board that clearly defines an artistic –
211
00:17:24.152 --> 00:17:33.137
area I want to explore from a design sense, and then I use that as a jumping off point for my work, is that copyrightable?
212
00:17:33.417 --> 00:17:36.298
Part of it, depending exactly what that design board needs.
213
00:17:36.758 --> 00:17:40.020
The trick about copyright is you cannot protect ideas.
214
00:17:40.380 --> 00:17:43.282
You protect the way those ideas are expressed.
215
00:17:43.782 --> 00:17:54.712
So to the extent that this is at a very high level and it's merely ideas, and the same thing would be true with very simple text prompts, even, you know, under a more lenient standard of what's protectable.
216
00:17:54.752 --> 00:18:01.598
If you just have a very simple prompt or a very simple design idea, that's not protectable.
217
00:18:02.248 --> 00:18:04.048
The tricky part is where do you go?
218
00:18:04.088 --> 00:18:08.890
Where's the line between what's an idea and what's the specific expression of that idea?
219
00:18:09.610 --> 00:18:14.351
So something like a drawing tends to be pretty clear where you have a design board.
220
00:18:14.371 --> 00:18:17.752
You know, it might depend specifically on what type of things are in there.
221
00:18:18.332 --> 00:18:27.494
But I still think it's helpful to the extent that you can show that there was this human input that's shaping and determining what the ultimate output looks like.
222
00:18:27.914 --> 00:18:28.854
That's very interesting.
223
00:18:32.475 --> 00:18:35.821
And I know I'm sure this is very hard to pinpoint.
224
00:18:36.462 --> 00:18:42.150
Is there a threshold that you think defines that at all?
225
00:18:42.190 --> 00:18:47.819
You know, is there a yeah, is there a threshold that is a good rule of thumb, let's say?
226
00:18:48.282 --> 00:18:50.303
That is very tricky because, again, it depends.
227
00:18:50.363 --> 00:18:53.544
You can think, for example, you think about just a movie in general.
228
00:18:53.664 --> 00:18:58.706
There are general themes of movies, boy meets girl, meet cute, and everything that are merely ideas.
229
00:18:58.746 --> 00:19:04.609
But as you start to get more specific about plot elements, how they meet, what they say, what they do.
230
00:19:04.669 --> 00:19:09.531
So at the one end, you have a script with very specific words and expression and direction.
231
00:19:10.291 --> 00:19:12.212
At the other end, you have these broad ideas.
232
00:19:12.812 --> 00:19:27.939
It is very hard to say where you've crossed that line, but definitely would need to be something very specific to what you're doing and something more than just the mere idea of, you know, again, a boy goes into a cafe.
233
00:19:28.619 --> 00:19:29.940
sits down, that's pretty high.
234
00:19:29.980 --> 00:19:36.706
But if you get into certainly dialogue, you're getting into drawings that show how they do it, what they look like.
235
00:19:37.226 --> 00:19:41.750
That's sort of the expression, the text, the images, and a lot of that detail.
236
00:19:42.271 --> 00:19:52.299
But again, I want to remind you that it's important that even for unprotectable elements, the way that you stitch that together, the way you combine those elements, is it self-protectable?
237
00:19:52.540 --> 00:19:54.721
Oh, at a certain point, is there a sort of
238
00:19:55.782 --> 00:19:59.384
Level of developed prompt where that prompt does become protectable.
239
00:19:59.404 --> 00:20:09.190
I mean, at a certain point, if I'm writing a prompt, there's a point at which that prompt becomes an essay and that essay becomes a thesis paper and that thesis paper becomes a novel.
240
00:20:09.210 --> 00:20:17.215
It's interesting about the Copyright Office view on this so far, because they said even if the prompts themselves are detailed enough to be protectable.
241
00:20:18.075 --> 00:20:27.722
somehow they're viewing those prompts as direction, as if, for example, if you hire somebody and you give them instructions as opposed to the content.
242
00:20:27.742 --> 00:20:36.688
Now, I would make a different argument because I think, again, a very simple prompt, save an image, you know, give me an image of a teddy bear sitting on a blanket.
243
00:20:36.949 --> 00:20:37.849
That's not protectable.
244
00:20:38.009 --> 00:20:45.455
As you mentioned, if you start to get very detailed and you start to describe in very granular detail exactly what that should look like,
245
00:20:46.035 --> 00:20:48.996
I think there are arguments that that should be sufficient human authorship.
246
00:20:49.456 --> 00:20:55.797
But in some of the initial rulings that the copyright office made, some of you may be familiar with the Zarya of the Dawn graphic novel.
247
00:20:56.217 --> 00:20:59.078
This was the first decision that the copyright office made.
248
00:20:59.558 --> 00:21:05.380
And they did show that they were very, very, you know, it wasn't just, you know, you put in a prompt and out came a picture.
249
00:21:05.400 --> 00:21:06.400
They regenerated.
250
00:21:07.080 --> 00:21:09.981
And these prompts, some of them were 100 words or more.
251
00:21:10.461 --> 00:21:12.161
And they made the argument.
252
00:21:12.241 --> 00:21:15.882
I thought it was a pretty convincing argument that it was detailed enough that it was
253
00:21:16.102 --> 00:21:18.244
to really shape and control the output.
254
00:21:18.284 --> 00:21:22.667
Because that's what the Copyright Office seems to want, is that it controls the output.
255
00:21:23.187 --> 00:21:25.389
Although in other areas, we don't really see that.
256
00:21:25.489 --> 00:21:29.752
And other rulings outside of AI, which is why I say I think they're a little hostile to AI.
257
00:21:30.473 --> 00:21:33.595
But where you get detailed enough, I think you should have arguments.
258
00:21:33.715 --> 00:21:41.541
And I mentioned there was another ruling which involved an image that actually won a prize at the Colorado State Fair,
259
00:21:42.181 --> 00:21:49.585
They also got this ruling that, you know, even though they generated and regenerated and had very detailed prompts, that wasn't protectable.
260
00:21:49.605 --> 00:21:51.486
And they've challenged this now in a court case.
261
00:21:52.006 --> 00:21:59.370
So I'm hoping that, you know, hopefully by the end of the year, because it's all been fully briefed, some of the initial arguments, that we'll have a decision.
262
00:21:59.430 --> 00:22:09.015
And I think I would not be surprised to see a court say that at least a very, very detailed prompt, because you could have a prompt that was so detailed that it is protected.
263
00:22:09.215 --> 00:22:12.858
really controlling every aspect of what that outcome is like.
264
00:22:13.078 --> 00:22:13.399
Yeah.
265
00:22:13.919 --> 00:22:23.888
But we'll have to see what, you know, what the court says, because the Copyright Office so far has just, you know, had a pretty clear direction that they don't think prompts alone are sufficient.
266
00:22:24.068 --> 00:22:24.248
But...
267
00:22:24.548 --> 00:22:25.429
And that's a block.
268
00:22:25.509 --> 00:22:26.451
It's just their opinion.
269
00:22:26.731 --> 00:22:33.500
Does that mean I could have edited together a series of things, but I could not own the characters?
270
00:22:33.600 --> 00:22:38.987
So I can own the final video if I've edited together lots of things because that's my human authorship.
271
00:22:39.167 --> 00:22:42.612
But I don't actually own any of the specific elements within that edit.
272
00:22:42.879 --> 00:22:57.773
Yeah, so if we assume that the individual elements, so let's say a character that was AI generated, no modifications, and again, this is the Copyary Office view, I don't know that they're right, but if they assume they're right, and a particular character or a particular element isn't protectable,
273
00:22:58.453 --> 00:23:03.794
If you select which ones you want to use and how you arrange them, that work as a whole.
274
00:23:03.854 --> 00:23:06.295
So that selection and arrangement is protectable.
275
00:23:06.875 --> 00:23:11.396
But that doesn't protect, as you noted, the individual unprotectable elements.
276
00:23:11.416 --> 00:23:12.917
You don't suddenly make them protectable.
277
00:23:12.937 --> 00:23:20.379
So if that's the case and those particular characters weren't protected, theoretically, you could take those individually.
278
00:23:20.419 --> 00:23:24.640
And if you could figure out what isn't protectable, you could take that and use it.
279
00:23:25.160 --> 00:23:37.654
And that's why when we advise clients, we say, you know, if you generate something, particularly through a text prompt, first of all, use those expressive prompts because then those characters will have your content in them and there won't be as much of a question.
280
00:23:38.274 --> 00:23:43.080
Let's say you take a very simple text prompt and now it comes a very simple character.
281
00:23:43.580 --> 00:23:47.642
Take that, modify it, because the modified version will be protectable.
282
00:23:47.722 --> 00:23:53.286
Now, the original version won't be, but a user isn't going to know what was the original version look like.
283
00:23:53.326 --> 00:23:54.226
What did you modify?
284
00:23:54.286 --> 00:23:57.728
So as a practical matter, it would be hard for them to figure that out.
285
00:23:57.808 --> 00:24:08.074
But if they could, yes, they could take out those unprotectable elements individually and use them, as long as they're not typing the way that you select it and arrange them into that final.
286
00:24:08.434 --> 00:24:09.335
In something like...
287
00:24:10.261 --> 00:24:16.888
James Bond, for example, James Bond himself, if that weren't protected, anybody could be making a James Bond film.
288
00:24:16.988 --> 00:24:23.014
And so it's all well and good to have owned Casino Royale or Goldfinger, whatever the first Bond film is.
289
00:24:23.434 --> 00:24:30.041
But if you don't own all the underlying characters enough, you've created one great thing, but everything else is free for all.
290
00:24:30.622 --> 00:24:38.293
Yeah, and there are public domain characters that people will write a work around, and they don't have the exclusive right to those characters.
291
00:24:38.453 --> 00:24:46.025
But they do have, you know, when they put a script together, when they put, you know, when they film and they have very specific elements in traditional media…
292
00:24:46.365 --> 00:24:47.407
that would be protectable.
293
00:24:47.447 --> 00:24:48.909
So it's somewhat analogous.
294
00:24:49.370 --> 00:24:57.942
You know, here you have the additional layer that, you know, when you're filming, when you're generating those images, there's this question mark about to what extent.
295
00:24:58.103 --> 00:25:02.329
But that's why we encourage people to make as many modifications as possible.
296
00:25:02.953 --> 00:25:09.897
And to insert in your prompting process, first of all, even though we don't know if it's protectable, make those prompts as detailed as possible.
297
00:25:09.977 --> 00:25:14.319
Make the argument that you are truly shaping what that final product is.
298
00:25:14.719 --> 00:25:18.401
As much as possible, add those sort of expressive prompts in addition.
299
00:25:19.121 --> 00:25:20.742
And I would say most of my clients do that.
300
00:25:21.162 --> 00:25:25.063
You've mentioned some of the decisions that have been coming down the pike about all of this.
301
00:25:25.143 --> 00:25:30.264
When I first saw you talk, there hadn't been any decisions on fair use in general.
302
00:25:30.304 --> 00:25:31.805
The Anthropic case hadn't occurred.
303
00:25:32.485 --> 00:25:40.427
It seems to me when all these things happen, on the one hand, you get answers and guidance for, this is the way everybody sees things now.
304
00:25:40.487 --> 00:25:44.227
And then very often it just raises a whole bunch more questions.
305
00:25:44.247 --> 00:25:50.829
Can you give us an update on where we are regulatorily regarding AI, copyright, and fair use?
306
00:25:51.169 --> 00:25:52.710
Yeah, this is the other big question, right?
307
00:25:52.730 --> 00:25:56.754
Which is one question is do I only the other question is if I use this, am I going to infringe somebody's right?
308
00:25:57.254 --> 00:26:01.798
And, you know, I've spoken on this topic and AI and copyright for, you know, over a dozen years.
309
00:26:01.858 --> 00:26:04.260
And for a long time, I had to say, well, there are no cases.
310
00:26:04.741 --> 00:26:07.863
And we've kind of looked at some analogous cases and all those cases.
311
00:26:07.884 --> 00:26:14.489
But we do have a couple of cases now or, you know, three to be exact and two that specifically deal with generative AI.
312
00:26:14.669 --> 00:26:14.789
Yeah.
313
00:26:15.550 --> 00:26:19.152
So we have a little bit of information, but sometimes it just raises more questions.
314
00:26:19.352 --> 00:26:22.713
And these two cases came out within three days of each other last year.
315
00:26:23.234 --> 00:26:27.216
And that's the Barts versus Anthropoc case and the Cadre versus Meta case.
316
00:26:27.816 --> 00:26:35.900
And just to set the stage on what the question is, as you mentioned fair use, and for people who may not be familiar with the concept, there's a lot of misunderstanding about fair use.
317
00:26:36.720 --> 00:26:48.113
Fair use is a concept that certain types of uses are permissible for things like criticism, comment, news reporting, etc., such that that use is not copyright infringement.
318
00:26:48.553 --> 00:26:50.775
And it's a balancing test, which is very difficult.
319
00:26:50.795 --> 00:26:56.201
It's very fact-specific, and I won't get into a lot of details, but there's four different factors you look at.
320
00:26:56.942 --> 00:26:59.964
And there's two particularly important factors that I will mention.
321
00:27:00.004 --> 00:27:06.508
The first one is the nature and purpose of work and whether it is what they call transformative.
322
00:27:07.408 --> 00:27:13.852
And that means, does it really just function for the same purpose as the underlying work, or does it add something new?
323
00:27:13.892 --> 00:27:18.195
Does it create something that is of a different nature or purpose or character?
324
00:27:18.215 --> 00:27:18.455
Yeah.
325
00:27:18.623 --> 00:27:24.973
The other factor that's important in particular is the impact of this use on the market for the work.
326
00:27:25.013 --> 00:27:27.277
So does it really harm the market for the work?
327
00:27:28.177 --> 00:27:36.000
And so when you do a fair use analysis, you're looking at those and the other two factors, which are the amount that you use and the nature of the work.
328
00:27:36.320 --> 00:27:45.024
So in the BARTs case, which is the first one to come out, well, I should say in both these cases, they found fair use on the specific facts.
329
00:27:45.664 --> 00:27:54.948
But that doesn't mean that this issue is resolved because there were a lot of limitations and caveats in these cases that give us a lot of pause.
330
00:27:55.288 --> 00:28:05.698
But one of the important things in the similarities in these cases is that both cases found that AI use is very transformative, which is important because that's one of the big factors.
331
00:28:05.738 --> 00:28:09.842
And in fact, one of the courts said it's probably the most transformative thing in our lifetime.
332
00:28:09.862 --> 00:28:10.282
Mm-hmm.
333
00:28:10.362 --> 00:28:14.463
And the idea is that you're using this because you're not, is that you're making copies.
334
00:28:14.923 --> 00:28:21.404
And the purpose of those copies is not to make copies, but to train, in a way, we're talking training.
335
00:28:22.404 --> 00:28:31.826
And I mentioned both these cases are focused on the training side, which is the copies that you make to create the tools, not output, which is a different issue when you output it.
336
00:28:32.126 --> 00:28:37.227
But in these cases, they said this was highly transformative and both found fair use on the facts.
337
00:28:37.747 --> 00:28:45.432
But in the BART's case, the distinction they made, they said, well, training use is okay, but they had two issues with what Anthropic had done.
338
00:28:46.213 --> 00:28:52.818
One is that they said that when they created this training database, they copied from pirated sites.
339
00:28:52.858 --> 00:28:57.221
So they didn't have, not just not unlicensed, but these were pirated sites.
340
00:28:57.361 --> 00:29:04.746
And that they also used this not just for training, but they created a central library of works that they used for training and other purposes.
341
00:29:04.766 --> 00:29:04.906
Yeah.
342
00:29:05.546 --> 00:29:10.708
So because of that, they found that those particular uses were not fair use.
343
00:29:11.008 --> 00:29:11.649
And that's where we.
344
00:29:11.789 --> 00:29:14.690
They hadn't actually paid for the work itself.
345
00:29:15.270 --> 00:29:23.773
And so that was considered theft because, you know, just like when we used to rent movies at Blockbuster, there was that FBI thing.
346
00:29:23.793 --> 00:29:25.434
You know, you can't share this.
347
00:29:25.494 --> 00:29:27.355
You can't publicly exhibit it.
348
00:29:27.555 --> 00:29:27.975
Not quite.
349
00:29:27.995 --> 00:29:33.197
And this is where the nuance is in because they said it was pirated, which is not the same as unlicensed.
350
00:29:33.377 --> 00:29:37.558
If you bought a book and you use that to train, that's not pirated.
351
00:29:37.658 --> 00:29:41.379
It's important to note that doesn't mean you were licensed to use it for training, but it wasn't pirated.
352
00:29:41.639 --> 00:29:44.000
And there's a question about what pirated means.
353
00:29:44.600 --> 00:29:47.701
Let's say you access it lawfully, but you didn't pay for it.
354
00:29:47.721 --> 00:29:52.582
So there's still a lot of questions about what pirated limits.
355
00:29:52.763 --> 00:29:57.704
But at least if you go to, because they went to these really well-known pirate websites that have pirated content.
356
00:29:58.144 --> 00:30:03.088
The other piece of it was the fact that they used it not just for training, but for other purposes.
357
00:30:03.609 --> 00:30:07.572
And there's a question about, well, if it was used just for training, would that have been okay?
358
00:30:08.453 --> 00:30:11.596
But the case did settle for, you know, $1.5 billion.
359
00:30:11.616 --> 00:30:17.101
And that's where you got a lot of things that said, oh, now if you don't license content for training, you're indulging.
360
00:30:17.141 --> 00:30:18.842
But again, it didn't say you have to license it.
361
00:30:18.882 --> 00:30:22.565
It just said really it should be non-pirated and used only for training.
362
00:30:22.585 --> 00:30:22.866
Right.
363
00:30:22.886 --> 00:30:23.066
Right.
364
00:30:23.688 --> 00:30:27.950
But the interesting thing is, so there, you know, that judge, that decision.
365
00:30:27.990 --> 00:30:34.194
But in the second case, the cadre case, the judge commented that he didn't care if it was pirated, that this was fair use.
366
00:30:34.274 --> 00:30:34.914
It didn't matter.
367
00:30:34.934 --> 00:30:36.575
So they disagreed on that point.
368
00:30:37.156 --> 00:30:43.940
At least, you know, one judge has said that it does matter how and where you got the content, at least if you use it for things other than training.
369
00:30:44.720 --> 00:30:49.782
But in the Quadrate decision, which was the second decision, again, he didn't care about the pirated issue.
370
00:30:50.342 --> 00:30:54.643
But his concern was in that last factor, which was the market harm.
371
00:30:55.304 --> 00:31:08.528
And he recognized a theory that sort of it's called indirect market substitution, which is the fact that if you flood the market with AI generated works, that might harm the market, even if the works individually aren't infringing.
372
00:31:09.108 --> 00:31:12.953
Now, in this case, he said he didn't find that because they didn't prove their case.
373
00:31:12.993 --> 00:31:19.761
But he suggested that others could and kind of gave a little road map of sorts of how others might make that claim.
374
00:31:20.022 --> 00:31:26.149
And this is the principle that it's fine for me to take a book, be inspired by it and write it in another book.
375
00:31:26.710 --> 00:31:33.895
But when an AI does it, it can potentially be writing 100 other books that are similar to that book at the same time.
376
00:31:34.256 --> 00:31:42.121
And that competes with me as a writer who I have a finite ability to output books on my own as a writer.
377
00:31:42.422 --> 00:31:46.485
But AI has a theoretically infinite ability to compete against that writer.
378
00:31:46.505 --> 00:31:46.905
Is that right?
379
00:31:47.445 --> 00:31:53.069
Yeah, but it's different because there's tool companies will make the argument that it's the analogy is the same.
380
00:31:53.129 --> 00:31:56.092
And so this argument will know it's different just because of the scale.
381
00:31:56.712 --> 00:32:03.234
But just like the cadre judge disagreed with the Barks judge, the Barks judge disagreed with the cadre case.
382
00:32:03.254 --> 00:32:09.036
Just this indirect market substitution theory had been raised by the Copyright Office and said, no, that that doesn't matter.
383
00:32:09.076 --> 00:32:10.237
That's just competition.
384
00:32:10.397 --> 00:32:18.099
So, again, we have two different cases that came out within a few days of each other, both of which had very, very different reasoning.
385
00:32:18.199 --> 00:32:19.840
So it's going to take a little more.
386
00:32:19.920 --> 00:32:22.161
And there's, you know, 100 more cases to follow.
387
00:32:22.801 --> 00:32:26.743
So it's going to be a little bit more to shake out exactly what those parameters.
388
00:32:26.803 --> 00:32:28.943
Does it matter if the content is pirated?
389
00:32:29.624 --> 00:32:36.886
Does it matter, you know, if market substitution, you know, is indirect and the fact that, you know, just the mass scale of AI?
390
00:32:37.066 --> 00:32:38.847
And so bottom line is we don't know.
391
00:32:38.907 --> 00:32:43.489
But one thing I also want to mention is that this involves the training piece.
392
00:32:44.249 --> 00:32:51.537
And that really doesn't directly impact users because the training issue has to do with how they train it.
393
00:32:51.977 --> 00:32:57.143
There are a number of cases that were coming down the pipe around output, but we haven't had decisions.
394
00:32:57.803 --> 00:33:06.813
And that is going to really be a matter of whether your output is substantially similar to some other work, likely something that was in that training data.
395
00:33:07.271 --> 00:33:28.293
Warner Brothers, Disney have sued Mid Journey along with NBC Universal because they recognize that if you at that point in time had prompted Mid Journey to show you Darth Vader, it would show you a character substantially similar to Darth Vader that was clearly referencing their own character.
396
00:33:28.354 --> 00:33:28.694
Is that right?
397
00:33:28.714 --> 00:33:28.814
Yeah.
398
00:33:29.015 --> 00:33:42.159
Yes, but a lot of just a lot of those cases are still about the training and they're showing examples of how the training, how the output shows up, you know, the prompt to show that the training, you know, that the training is in fair use.
399
00:33:42.819 --> 00:33:46.260
Some of the cases, not all of them also address output.
400
00:33:47.200 --> 00:33:51.361
But a lot of these, a lot of the cases are focused really solely on the training piece.
401
00:33:51.501 --> 00:33:54.962
And all of the cases so far have only been against tool providers.
402
00:33:55.383 --> 00:34:06.526
We don't yet have a case that's really been brought against a studio or somebody that's using AI as a tool that's not training models, but that is using that as output.
403
00:34:06.966 --> 00:34:15.969
And that's, you know, the other, you know, sort of practical advice that I'm often working with clients is as a user of an AI tool, you don't generally and, you know, all of this is still.
404
00:34:16.289 --> 00:34:22.271
You're OK. Not in the law, but you're not going to generally it's unlikely that you're going to be responsible for the training.
405
00:34:22.291 --> 00:34:25.472
What you need to worry about is making sure your output isn't similar.
406
00:34:25.612 --> 00:34:35.455
And you have a lot of control over that because a lot of these cases where they gave examples, those are highly orchestrated examples intended to generate similar content.
407
00:34:35.815 --> 00:34:45.160
prompting things and trying to emulate something, you know, give me this, you know, in the style of, you know, this artist, even though style itself isn't actually protectable.
408
00:34:45.720 --> 00:34:56.706
If you reference IP in your prompts, you have a greater risk that you will somehow, even inadvertently, end up with output that might be too close to another work.
409
00:34:56.726 --> 00:35:03.830
So if you avoid doing that, well, first of all, if you start with those expressive prompts, drawings and things that you create,
410
00:35:04.542 --> 00:35:14.796
As a practical and as a technical matter, that's likely to dominate the output in a way that it's unlikely that inadvertently some random work and the training data is going to show up in your output.
411
00:35:15.368 --> 00:35:34.407
Even if you, you know, and if you prompt things in a way that's generic and you're not trying to emulate and get output that looks like someone else's, I just think that the practical risk that you're going to out of all the images that we're trained on, that you're going to pull up something that happens to be substantially similar to a work that was in the training data.
412
00:35:35.375 --> 00:35:36.376
is not as likely.
413
00:35:36.716 --> 00:35:43.382
And I'll also mention that there are things that you can do as a user, too, to try to limit that risk when you choose tools that you use.
414
00:35:44.323 --> 00:35:50.069
You know, there are tools, you know, most of these tools, at least in an enterprise version, also have indemnities around use.
415
00:35:50.149 --> 00:35:54.232
Now, those indemnities have limits and caveats and restrictions.
416
00:35:54.313 --> 00:35:58.617
And so you need to look at that and make sure that you, you know, we're kind of following that.
417
00:35:59.277 --> 00:36:02.059
And that can be something when you're choosing tools to look at.
418
00:36:02.599 --> 00:36:20.970
But I think, you know, users, while we can't say that, you know, in an uncertain legal environment that the risk is zero, I think that the practical risk, if you're not trying to have the output look like something else and you take reasonable measures to avoid that, the risk that you're going to end up with something that somehow infringes inadvertently is true.
419
00:36:21.671 --> 00:36:27.433
You know, so each of these decisions has sort of moved the conversations forward, brought up some other questions.
420
00:36:28.213 --> 00:36:34.495
But can we fairly confidently say that the fair use argument has generally won?
421
00:36:34.575 --> 00:36:37.356
I mean, not without saying that's not OK, that's not good.
422
00:36:37.376 --> 00:36:43.698
But the central fair use argument, the training on, you know, copyright data is –
423
00:36:44.478 --> 00:36:45.359
sort of okay?
424
00:36:45.459 --> 00:36:49.502
I mean, has that been part of the argument been generally resolved?
425
00:36:49.862 --> 00:36:57.648
It's too early to say that because what we, you know, we have two cases, first of all, only two cases, and you can have different circuits that disagree.
426
00:36:57.668 --> 00:36:59.609
You can have other cases that are totally different.
427
00:36:59.629 --> 00:37:02.091
So it's far too early to draw conclusions.
428
00:37:02.511 --> 00:37:10.257
And because, as I mentioned, it's such a fact-specific analysis that the fact of each case can also mean that in one situation, something's fair use.
429
00:37:10.797 --> 00:37:10.998
Yeah.
430
00:37:11.018 --> 00:37:11.718
And not.
431
00:37:12.279 --> 00:37:12.639
It's not.
432
00:37:12.699 --> 00:37:20.988
Now, there have been some beneficial in these first two rulings, some beneficial things, one of which is this transformative factor, because that is very important.
433
00:37:21.709 --> 00:37:26.514
And the fact that both of these cases said it's highly transformative is very helpful.
434
00:37:26.554 --> 00:37:32.241
The other thing that was very helpful is that in looking at that last factor, which is the impact of the market.
435
00:37:32.821 --> 00:37:41.349
Both of these cases said that what you don't look at is the market for licensing the work itself, because there's an argument that that will make.
436
00:37:41.389 --> 00:37:48.116
And it's kind of I've always thought it was very circular that, you know, if you say, you know, if I had licensed this work,
437
00:37:48.536 --> 00:37:52.258
these materials to you for training, I've now lost that income.
438
00:37:52.578 --> 00:37:53.799
And so that's a good harm.
439
00:37:53.919 --> 00:37:58.521
But that's exactly the harm you're trying to figure out if you need to license.
440
00:37:58.561 --> 00:38:00.702
So it's always a bit circular to count that.
441
00:38:01.323 --> 00:38:09.747
And both of these cases did say, no, you don't look at the loss of marketing from licensing the thing we're trying to decide if you have to license.
442
00:38:10.307 --> 00:38:14.029
So I think these cases both had some beneficial rulings, but because...
443
00:38:15.590 --> 00:38:20.513
You know, in the BART's case, you had this question of you still have to look at where the content comes from.
444
00:38:21.113 --> 00:38:25.016
On CADRE, you still had this question of, you know, indirect market substitution.
445
00:38:25.056 --> 00:38:32.600
From that, you know, there were enough caveats in that that until we have a little more case law coming out of this, you know, we're really not going to know.
446
00:38:32.620 --> 00:38:36.203
And I frankly think it's going to be, you know, quite a while because there's so many cases.
447
00:38:36.283 --> 00:38:42.346
Each could come out differently, even have different circuits, different parts of the country might rule differently.
448
00:38:42.406 --> 00:38:44.447
And these are lower court decisions, too.
449
00:38:45.368 --> 00:38:54.298
Again, we're not very close to resolving the core issue, which is, you know, is training on copyrighted material fair use?
450
00:38:54.759 --> 00:38:57.522
We don't really have a clear answer.
451
00:38:57.542 --> 00:39:00.205
So a few early wins, but the game isn't over yet.
452
00:39:00.802 --> 00:39:04.826
Yeah, and I'd say a few early wins, but a few early caveated limited wins.
453
00:39:04.866 --> 00:39:07.409
Right, okay.
454
00:39:07.449 --> 00:39:12.914
Just because they didn't fully win, I'm sure that won't stop the tech companies from declaring victory.
455
00:39:13.215 --> 00:39:15.117
Does it feel a little bit like the...
456
00:39:16.553 --> 00:39:36.010
creator, the end creator, if they're an individual and not an enterprise or not a brand, is generally protect, seems to be at the moment, whether it's just because they're too small fry or not, generally seems to be pretty protected in terms of their freedom to use these things, so long as they're not monetizing something that is potentially infringing.
457
00:39:36.390 --> 00:39:40.674
But they sort of have a certain degree of freedom to create, at least.
458
00:39:42.097 --> 00:39:48.742
One of the factors is, you know, if it's kind of a nonprofit use versus for profit, but that's really not a key factor.
459
00:39:49.283 --> 00:40:02.934
I think for people who are using those creatives who are creating materials, there is, you know, if you are outputting as part of this something that is substantially similar to an existing work, there is risk.
460
00:40:03.635 --> 00:40:07.297
But my advice to them is to take steps to make sure that you're not doing that.
461
00:40:08.138 --> 00:40:14.661
One of the things that I keep coming back to is there's an online creator who's done something called Star Wars Beggars Canyon.
462
00:40:15.041 --> 00:40:21.124
And it's an AI piece that's a Luke Skywalker story, I think, in between, you know, some of the films.
463
00:40:21.585 --> 00:40:23.425
And he's sort of decided to take it upon himself.
464
00:40:23.646 --> 00:40:28.828
I've been shocked that Disney hasn't sent a cease and desist or hasn't pressured YouTube to stop it.
465
00:40:29.468 --> 00:40:31.769
And I think it's because people really quite like it.
466
00:40:32.030 --> 00:40:36.492
I think it's, you know, at a time where, you know, the Star Wars film,
467
00:40:37.232 --> 00:40:42.714
franchise is losing a lot of brand relevance, this is gaining quite a lot.
468
00:40:42.814 --> 00:40:52.999
And I think that there's a degree to which, so long as people aren't, you know, he's not going to a movie theater and exhibiting it around his town, but he's just putting it on YouTube.
469
00:40:53.539 --> 00:40:54.880
I think that there's a degree to which
470
00:40:55.680 --> 00:40:58.083
Certain creators are being allowed to just go.
471
00:40:58.183 --> 00:41:00.425
But like you said, it's not without risk.
472
00:41:00.685 --> 00:41:05.710
He is still at risk of them sending a cease and desist and taking that hard work that he's done down.
473
00:41:05.870 --> 00:41:06.391
Yeah.
474
00:41:06.451 --> 00:41:08.493
And that helps me frame a little bit more what you're asking about.
475
00:41:08.513 --> 00:41:12.337
So fan fiction is always an interesting area for companies in terms of policing itself.
476
00:41:12.777 --> 00:41:15.520
Many people know Disney tends to be very aggressive in policing things.
477
00:41:15.560 --> 00:41:17.862
So it is interesting their view here.
478
00:41:17.882 --> 00:41:25.829
But with fan fiction, you know, a company who's a rights holder has to determine, you know, whether this is something that's beneficial or not.
479
00:41:26.830 --> 00:41:38.079
And this is where when you get into this, again, fair use, you know, all these factors, if it is not, you know, if it's not monetized, if it's not for profit, that is one part of the fair use analysis.
480
00:41:38.099 --> 00:41:38.199
Yeah.
481
00:41:38.980 --> 00:41:46.787
But, you know, things that aren't for profit, if you're still, you know, if you're still copying material, taking content, you know, there's still going to be arguments for infringing.
482
00:41:46.807 --> 00:41:51.931
And I say arguments because, again, you have depends on how much you use, how you use it, so many factors.
483
00:41:52.491 --> 00:41:59.377
But I think ultimately the decision of a company of whether to go after people who are doing fan fiction is a practical risk.
484
00:41:59.397 --> 00:42:01.839
So it's not that it isn't infringing necessarily.
485
00:42:02.380 --> 00:42:05.723
It may be that they think it benefits the brand so that they tolerate it.
486
00:42:07.884 --> 00:42:13.489
But it's a tricky for brands and IP owners.
487
00:42:14.289 --> 00:42:22.636
It's a delicate balance because if you want to make sure if you don't enforce it, is that going to affect ability to enforce for other types of uses you want to make?
488
00:42:22.656 --> 00:42:22.996
Right.
489
00:42:23.196 --> 00:42:28.880
But fan fiction has always been, you know, even outside of anything with AI has always been an area where...
490
00:42:29.761 --> 00:42:38.287
Content owners are very careful about what they do because they do want to, they can be very beneficial to have this, but it can cross a line too.
491
00:42:38.547 --> 00:42:45.232
Because before AI came out, it felt like every big brand and media company was obsessed with figuring out user-generated content.
492
00:42:45.352 --> 00:42:47.414
They were like, great, we want to give them assets.
493
00:42:47.454 --> 00:42:48.575
We want them to post online.
494
00:42:48.996 --> 00:42:51.578
And then suddenly AI allowed them to do it at scale.
495
00:42:51.979 --> 00:42:56.543
And suddenly, you know, suddenly the big media companies became very nervous about UGC.
496
00:42:57.144 --> 00:43:02.249
And, you know, and then the decision was sort of made for them because all of their stuff was in these models.
497
00:43:02.469 --> 00:43:04.371
It's been interesting to see them wrestle with that.
498
00:43:04.511 --> 00:43:09.698
Well, here's the other interesting thing is you have some people who are really in the middle on both sides.
499
00:43:09.718 --> 00:43:13.303
So you may be a content owner that's very interested in protecting your stuff.
500
00:43:14.024 --> 00:43:21.554
At the same time, particularly studios, you may be very interested in using AI and all the efficiencies and other things that it can bring.
501
00:43:21.574 --> 00:43:21.834
So...
502
00:43:23.695 --> 00:43:26.696
You're in this middle because you're not just a content owner.
503
00:43:26.716 --> 00:43:28.877
You're also a content creator that wants to use it.
504
00:43:28.917 --> 00:43:33.878
And so it's been interesting to see the views of the studios change.
505
00:43:33.958 --> 00:43:41.541
And I think although we still have all these cases on training, a lot of the conversations I'm hearing are really focusing on that output question.
506
00:43:41.601 --> 00:43:44.862
Because there may be, like, obviously can't speak for the studios.
507
00:43:44.942 --> 00:43:45.682
I don't worry about it.
508
00:43:45.982 --> 00:43:57.305
Not speaking on behalf of anybody, myself, but I think our questions about the training, again, so these cases, but, you know, the output and making sure that you're protecting.
509
00:43:57.385 --> 00:44:07.667
So, you know, Disney, that people aren't generating stuff that looks like their content is a far more critical issue than the training issue.
510
00:44:07.687 --> 00:44:08.247
Yeah.
511
00:44:08.507 --> 00:44:18.232
I'm not saying they're giving up on the training issues, but I think I'm just hearing conversations really focusing a lot on output and less on training than I did say a year ago.
512
00:44:18.793 --> 00:44:24.176
Obviously, there's a lot of fear around professional creators feeling like there are so many people coming into it.
513
00:44:24.556 --> 00:44:29.319
There's a lot of concern around the fair use argument and job displacement and things like that.
514
00:44:29.679 --> 00:44:32.541
Should people be thinking about this legally?
515
00:44:32.601 --> 00:44:36.963
Are there basic things that they can be doing to protect themselves and,
516
00:44:37.103 --> 00:44:52.622
And their work, even at the most base level, is it possible to protect this thing on my own or our must I be engaging with somebody like you who's a, you know, a copyright expert and a lawyer to help shepherd me through these protections?
517
00:44:52.963 --> 00:44:59.605
Let me start with the beginning of your comments, because it touched on something we haven't really talked about, which is the job display.
518
00:44:59.665 --> 00:45:01.625
And so it's not because we've talked about copyright.
519
00:45:02.305 --> 00:45:16.609
But one of the other issues that I talk about a lot and focus on are the sort of digital replicas and this concern about talent, because there's a lot of concern among talent that AI is going to replace them.
520
00:45:23.330 --> 00:45:25.392
He's going to be a guest in just a couple of weeks.
521
00:45:25.712 --> 00:45:25.992
Yeah.
522
00:45:26.133 --> 00:45:26.733
Duncan's great.
523
00:45:26.793 --> 00:45:27.514
I love Duncan.
524
00:45:28.354 --> 00:45:33.618
And there's obviously a lot of legitimate concerns that creators, excuse me, that talent has.
525
00:45:34.119 --> 00:45:39.283
But I'd also say that technology has been replacing people for years.
526
00:45:39.723 --> 00:45:42.986
CGI, we used to hand animate things, right?
527
00:45:43.686 --> 00:45:45.648
We used to hand screw bolts on cars.
528
00:45:45.708 --> 00:45:47.830
Technology has always replaced jobs.
529
00:45:48.430 --> 00:45:49.571
And this isn't going away.
530
00:45:49.851 --> 00:45:50.051
Yeah.
531
00:45:50.091 --> 00:45:52.053
Nobody gets their start in the mailroom anymore.
532
00:45:52.393 --> 00:45:52.633
Yeah.
533
00:45:52.813 --> 00:45:54.994
And there's benefits to AI, even for creators.
534
00:45:55.014 --> 00:46:00.755
You can also, if you create a digital, you may be able to make money without having to go in and spend a day of work.
535
00:46:00.855 --> 00:46:05.056
And I know that you all work very hard to make sure that people get compensated and have consent.
536
00:46:05.536 --> 00:46:07.997
But I do think that there's a lot of fear of talent.
537
00:46:08.017 --> 00:46:11.358
But I think AI can also do something that...
538
00:46:11.998 --> 00:46:25.353
that's really beneficial for creators, which is, you know, reduce barriers to entry, you know, elevate diverse voices, because now you can create a film, you can create content without having to have as much money and have a studio behind you.
539
00:46:25.413 --> 00:46:27.235
So there are always trade-offs on this.
540
00:46:27.515 --> 00:46:29.697
And while I'm sensitive to job displacement,
541
00:46:30.398 --> 00:46:34.667
I also think that this is something we're just going to have to figure out because it's not going away.
542
00:46:35.208 --> 00:46:41.981
And, you know, we have, you know, if you use somebody's, there's, let's say there's protections if you're using somebody's replica.
543
00:46:42.749 --> 00:46:47.574
There's also synthetic talent, and the unions are trying to arrest that through rules.
544
00:46:48.035 --> 00:47:01.209
With synthetic talent, to go back to the meaningful human intervention, if that's all synthetic, even if it's sort of performance capture, is that performance capture considered a meaningful human intervention?
545
00:47:01.229 --> 00:47:01.449
No.
546
00:47:01.749 --> 00:47:03.149
outside of copyright land now.
547
00:47:03.189 --> 00:47:04.950
So we're not worried about originally.
548
00:47:04.970 --> 00:47:16.973
We're just talking about if a studio wants to, instead of hiring a real person, create the Tilly Norwood that got a lot of it and create a fully synthetic human being to be a performer.
549
00:47:17.033 --> 00:47:19.193
And then they don't, and they're not using someone's likeness.
550
00:47:19.213 --> 00:47:20.873
They don't have a variety of publicity issues.
551
00:47:21.354 --> 00:47:24.394
And theoretically, they can do that without paying talent.
552
00:47:24.834 --> 00:47:30.596
Now, what the union has done in their most recent contract, they're beginning to put provisions that require, if you're a union
553
00:47:32.076 --> 00:47:44.903
To make it as expensive to use synthetic talent as to use human talent by contractually requiring companies to only use synthetic talent if there's a good reason to do that and to have to make payments.
554
00:47:44.943 --> 00:47:46.084
So contractually.
555
00:47:46.164 --> 00:47:52.007
Is there an engineer at the end of the Tilly Norwood avatar that gets the payment?
556
00:47:52.395 --> 00:47:52.975
No, it's not.
557
00:47:53.015 --> 00:47:55.976
There really isn't because it's not.
558
00:47:56.196 --> 00:48:05.180
There are tools that can be used, but you're talking about rights, person rights, rights of publicity, rights to use somebody's name, image, likeness.
559
00:48:05.200 --> 00:48:06.280
There's no person there.
560
00:48:06.360 --> 00:48:10.482
So just like if you think about it in a video game, you create characters.
561
00:48:10.502 --> 00:48:11.682
They're not real people.
562
00:48:11.702 --> 00:48:13.763
You don't pay anybody because you created characters.
563
00:48:14.503 --> 00:48:21.824
This is a real, this is like creating a video character, except it's replicating and replacing what you might have.
564
00:48:21.845 --> 00:48:25.345
But if you think about that, so animation in a sense does, right?
565
00:48:25.405 --> 00:48:26.745
So it's not a lot of stuff.
566
00:48:27.006 --> 00:48:30.506
I guess my point is it's not really that different than we've done for many years.
567
00:48:30.586 --> 00:48:35.087
It's just suddenly the scale of this and the concern seems different.
568
00:48:35.127 --> 00:48:43.329
But I would argue that it's really no different than, you know, the synthesizer, the CGI, you know, factories that have, you know,
569
00:48:44.545 --> 00:48:55.112
Modernization does replace that, and we hope, and there's many who argue that even though it might reduce certain opportunities, it also creates volumes of opportunities and new opportunities.
570
00:48:55.152 --> 00:48:59.995
So there's a lot of arguments for pros and cons, and people have to decide their own comfort.
571
00:49:00.576 --> 00:49:08.105
with the, quote, morality and ethics around AI use, both in respect to copyright, but also in respect to things like digital replicas.
572
00:49:08.966 --> 00:49:15.635
But I still say this isn't all that different than stuff that's been going on for years, except that it's at a little bit more of a massive scale.
573
00:49:15.875 --> 00:49:30.126
Is there a concern for new actors, potentially, is that you could sort of be competing against Meryl Streep for the rest of your life after she moves on and her digital self is there still competing for roles?
574
00:49:30.667 --> 00:49:33.829
You're constantly being judged against the greatest who's ever done it.
575
00:49:33.889 --> 00:49:36.672
I hadn't thought about it that way, but I guess...
576
00:49:37.637 --> 00:49:39.038
I guess that's true to some extent.
577
00:49:39.738 --> 00:49:44.840
And this isn't, you know, obviously this is true across the board, not just in entertainment, but even lawyers.
578
00:49:44.900 --> 00:49:47.881
We're worried about lawyers replacing, you know, lawyers with all this.
579
00:49:47.961 --> 00:49:49.822
It's just something as a society that...
580
00:49:50.760 --> 00:49:52.141
you know, that we have to deal with.
581
00:49:52.301 --> 00:49:59.206
And, but I do think that, you know, there is, that it's not, this is my personal view, it's not going to replace real people.
582
00:49:59.246 --> 00:50:08.051
People, there's a certain warmth and depth to a performance that at least at this point, I don't think you see from a lot of these AI generated performances.
583
00:50:08.352 --> 00:50:11.654
And I think people will still want to connect with specific actors.
584
00:50:11.674 --> 00:50:19.439
So I don't think it's going to, you know, I don't think in 10 years or even in three, you know, three years, 10 years, we're going to suddenly have like no real actors and all talents going to be AI.
585
00:50:19.899 --> 00:50:22.882
But, you know, there definitely will be a role for that.
586
00:50:23.402 --> 00:50:30.167
And this will help reduce costs and make people be able to make films and do things in films that otherwise they couldn't do.
587
00:50:30.207 --> 00:50:37.933
Stunts, there's a lot of things that can be done more safely or scenes that would have just been ridiculously expensive to produce that now you can do.
588
00:50:38.073 --> 00:50:42.437
So, you know, there's always trade-offs when you're dealing with new technology.
589
00:50:43.157 --> 00:50:44.778
My thesis has always been that...
590
00:50:46.376 --> 00:50:57.769
The movies that we see, the 15 movies that Warner Brothers releases every year, the 15 that Paramount releases every year, they're still going to release those, but it is going to create a middle class of film.
591
00:50:57.789 --> 00:51:02.094
I mean, only 1% of all the screenplays that have ever been written have been made into movies.
592
00:51:02.534 --> 00:51:11.920
You've got an incredibly rich history of human creativity that's just sitting there that has never been activated, never been used, never been made into a film.
593
00:51:12.341 --> 00:51:14.342
And a lot of it's probably very good.
594
00:51:14.442 --> 00:51:18.485
I've read spec scripts that are fantastic that I'm fairly sure will never get made.
595
00:51:19.125 --> 00:51:21.167
AI sort of has the ability to...
596
00:51:21.787 --> 00:51:24.428
take a lot of that and make it active again.
597
00:51:24.528 --> 00:51:26.289
And I'm quite encouraged by that.
598
00:51:26.329 --> 00:51:27.990
I think that you're right.
599
00:51:28.050 --> 00:51:34.573
We still want to see people do things and human beings be expressive and we still want to connect with that humanity.
600
00:51:35.113 --> 00:51:45.698
But in terms of stories being told, I think there's a whole level of films that will be end-to-end AI, films that'll be hybrid productions, and then obviously the big budget productions.
601
00:51:45.738 --> 00:51:49.260
I just think it's just going to broaden the aperture.
602
00:51:49.647 --> 00:51:53.910
And we've seen that given even putting AI aside like this, we started to get into the internet.
603
00:51:53.930 --> 00:51:57.092
It used to be, you know, if you wanted entertainment, you had TV, you had movies.
604
00:51:57.452 --> 00:51:58.253
Well, now look at this.
605
00:51:58.293 --> 00:52:11.202
We have videos and, you know, young people today, they don't, not that they don't go to movies, but they spend a lot more time watching digital content and short form content, which did not exist, you know, 15 years ago or however many years ago.
606
00:52:11.602 --> 00:52:16.406
So that was already a shift that, and that has impacted, you know, film and television and all of this.
607
00:52:17.086 --> 00:52:22.231
But when we say that, we don't necessarily call that evil, but AI, there's definitely people who don't like that.
608
00:52:22.251 --> 00:52:23.752
But this is just the evolution.
609
00:52:24.232 --> 00:52:25.233
Content has evolved.
610
00:52:25.293 --> 00:52:27.596
People want to maybe have more interaction with content.
611
00:52:28.236 --> 00:52:34.141
And I think this is yet another evolution of how we interact and the fact that, you know, and the same thing in the music industry.
612
00:52:34.161 --> 00:52:37.004
It used to be you couldn't put out a music unless you were at a record company.
613
00:52:37.024 --> 00:52:40.187
Well, now you can do a lot of things without needing to have a record contract.
614
00:52:41.127 --> 00:52:47.930
You can become a big star without ever having a contract if you can put your content out on your own.
615
00:52:47.990 --> 00:52:51.911
So these are all shifts in this industry, major shifts.
616
00:52:52.631 --> 00:52:57.553
But that's the same old story that we've been seeing over time, technology.
617
00:52:57.573 --> 00:52:57.673
Yeah.
618
00:52:58.233 --> 00:52:59.993
and changes that can be monumental.
619
00:53:00.233 --> 00:53:05.555
And, you know, it's going to be hard to see, you know, what is the entertainment landscape going to look like in five years?
620
00:53:05.795 --> 00:53:10.836
You know, I think it's hard to predict, but I will predict that we won't see like no humans in content.
621
00:53:11.156 --> 00:53:13.997
Well, I don't think we're all AI generated.
622
00:53:14.037 --> 00:53:16.237
I think there will still be a big role for people.
623
00:53:16.717 --> 00:53:24.779
And most of my clients, at least, even though they use AI as a tool to assist them, they're still very big believers in the importance of the human contribution.
624
00:53:24.779 --> 00:53:28.481
And I think that people forget that creating a movie is fun.
625
00:53:29.001 --> 00:53:30.482
It's a form of entertainment itself.
626
00:53:30.942 --> 00:53:34.124
And I think that that's going to become much more a part of, I think, the discussion.
627
00:53:34.144 --> 00:53:38.126
I mean, Roblox has, you know, hundreds of millions of users.
628
00:53:38.447 --> 00:53:44.030
At least 50%, if not more, are the creators of the Roblox games.
629
00:53:44.470 --> 00:53:48.152
It's kids going on there to make a game that other people will play.
630
00:53:48.572 --> 00:53:50.213
And that's entertainment, too.
631
00:53:50.293 --> 00:53:52.374
That's not just a monetized engagement.
632
00:53:52.455 --> 00:53:53.015
That's fun.
633
00:53:53.035 --> 00:53:53.275
You should.
634
00:53:53.595 --> 00:53:58.897
Yeah, we shifted from this notion of entertainment as a passive thing to as something as a participatory thing.
635
00:53:58.917 --> 00:54:00.378
And I think that's continuing.
636
00:54:01.138 --> 00:54:02.879
You know, I don't think that's going away.
637
00:54:03.299 --> 00:54:06.701
I think many people can think that that's, you know, that that's a good thing.
638
00:54:07.141 --> 00:54:08.422
I think that's a great place to end.
639
00:54:08.882 --> 00:54:10.703
Thank you so much for doing this with me.
640
00:54:10.963 --> 00:54:15.205
I'm, you know, I'm always completely fascinated when you talk about all this.
641
00:54:16.586 --> 00:54:19.609
And so I'm so grateful that you came on the podcast.
642
00:54:19.809 --> 00:54:23.693
And I'm still just somebody learning on the other side of it.
643
00:54:23.753 --> 00:54:30.960
And I so appreciate how you talk about these things and the education certainly that I've gotten from you on it.
644
00:54:31.064 --> 00:54:31.885
Well, thanks for having me.
645
00:54:31.905 --> 00:54:32.766
I love talking about it.
646
00:54:32.806 --> 00:54:33.366
It's fun.
647
00:54:33.467 --> 00:54:34.287
It's entertaining.
648
00:54:34.387 --> 00:54:38.352
And I really do want to help get some good information out there because it's just so much confusion.
649
00:54:38.372 --> 00:54:40.474
So thank you for letting me share a little bit.
650
00:54:40.734 --> 00:54:41.355
Oh, it's a pleasure.
651
00:54:41.375 --> 00:54:45.279
Well, maybe when some of these other decisions come out, we'll have you back.
652
00:54:45.299 --> 00:54:45.880
Okay.
653
00:54:46.180 --> 00:54:46.901
Happy any time.
654
00:54:47.588 --> 00:54:47.909
All right.
655
00:54:47.949 --> 00:54:48.389
Terrific.
656
00:54:48.429 --> 00:54:49.230
Thank you so much.
657
00:54:49.891 --> 00:54:50.852
So she's great.
658
00:54:50.892 --> 00:54:52.393
She explains things so well.
659
00:54:52.814 --> 00:55:03.525
And it's so nice to have a conversation about copyright where you actually feel like you're getting somewhere and you can kind of create some footholds to give you a place to just start in understanding all of this.
660
00:55:04.165 --> 00:55:06.227
Because I think we throw around copyright a lot.
661
00:55:06.267 --> 00:55:10.712
And this reminded me of our chat with Keith Kufferschmid from the Copyright Alliance.
662
00:55:10.932 --> 00:55:12.333
These are incredibly important things.
663
00:55:12.513 --> 00:55:14.813
We're figuring out how to make them work in the workflow.
664
00:55:15.574 --> 00:55:16.634
And the answer isn't simple.
665
00:55:16.774 --> 00:55:19.895
And it depends on how much is used, what territory you're in.
666
00:55:20.176 --> 00:55:30.339
Organizations like SAG, and we had Duncan Crabtree Ireland, the national director of SAG on this podcast, and they worked very hard to make clearly defined rules around what's possible and what's not.
667
00:55:30.800 --> 00:55:33.881
In the UK, for example, equity hasn't taken a position.
668
00:55:34.021 --> 00:55:37.502
That just shows how slow these organizations are to act around these things.
669
00:55:37.962 --> 00:55:39.603
And really where the rubber meets the road is,
670
00:55:40.023 --> 00:55:41.525
is in these legal decisions.
671
00:55:41.625 --> 00:55:48.614
And people like Lisa, when she's called to do something like Critters, she's really helping figure out the case law on this.
672
00:55:48.654 --> 00:55:49.795
And she's really in the trenches.
673
00:55:49.835 --> 00:55:59.607
And this is why it's very important that you talk to your business affairs people, that you talk to your lawyers to make sure that you really understand how this works in your production workflows.
674
00:56:00.035 --> 00:56:01.175
So anyway, that was great.
675
00:56:01.215 --> 00:56:02.055
Thank you for listening.
676
00:56:02.296 --> 00:56:07.537
Next week, we're going to have Hisko Holsing, who is a fabulous Dutch filmmaker and artist.
677
00:56:08.197 --> 00:56:12.718
And I tracked him down because I saw his film get booed at the Annecy Film Festival.
678
00:56:13.298 --> 00:56:15.619
And I watched the film, and it was fantastic.
679
00:56:16.259 --> 00:56:20.500
And what became apparent in watching all of his making-of videos is...
680
00:56:21.060 --> 00:56:26.962
He oil painted these massive paintings and they were used as the storyboards.
681
00:56:27.342 --> 00:56:31.404
And then he made this film, did it with lots of artists in three different countries.
682
00:56:31.564 --> 00:56:34.545
And he trained a model on his own work.
683
00:56:34.685 --> 00:56:39.287
And that model helped give the final piece an oil painting-y look.
684
00:56:40.007 --> 00:56:41.668
And it was people walked out.
685
00:56:41.708 --> 00:56:42.768
They protested.
686
00:56:43.348 --> 00:56:46.770
And in terms of, it reminded me of Brady Corbet's
687
00:56:48.279 --> 00:56:49.099
the Brutalist film.
688
00:56:49.259 --> 00:56:54.821
AI used in a very small way, but just because somebody touched AI, they wanted to cancel him.
689
00:56:55.101 --> 00:56:56.541
I reached out to him on Instagram.
690
00:56:57.681 --> 00:56:58.862
We sat down for a call.
691
00:56:58.902 --> 00:57:02.143
He was quite emotional, actually, and I can understand why.
692
00:57:02.483 --> 00:57:11.905
The film that he was creating was about war and trying to say something about war, and he unexpectedly became the poster boy for using AI in film, and it's not what he wanted to be.
693
00:57:12.445 --> 00:57:15.492
But anyway, I'm very happy to be having that conversation with him.
694
00:57:16.575 --> 00:57:17.517
That will be next week.
695
00:57:17.637 --> 00:57:18.359
Please tune in.
696
00:57:18.439 --> 00:57:20.183
Thank you for listening, and we'll see you next week.