00:00
And we would like to welcome John Rizvi to the program. He's known as The Patent Professor. He's an adjunct professor of patent law at Nova Southeastern Law School in Florida. And he is the author of two books on patents. And this morning, we are going to talk about patents, copyrights and chatGPT, and AI. Good morning, John.
00:23
Good morning. Always a pleasure to be here.
So, John, this is getting, this is going very quickly, you know, the more that we hear about chat GPT, AI generated content, even artwork, and that sort of thing. What, in terms of patents, do we need to know about, and or copyright protections?
00:50
Yeah, so, I mean, the most important thing is, under the Constitution, Article One, Section Eight, intellectual property patents and copyrights are for limited times to authors and inventors. And that language is specifically in the Constitution. That's where we get the basis for all the intellectual property laws. And the question becomes, is a machine an author? Can a computer be an author of the work and that's, you know… I don't know, if you recall, several years ago, there was the monkey selfie case, where a monkey got a hold of a cell phone, took a bunch of selfies, the images became extremely valuable, and the owner of the cell phone tried to claim copyright. The court held that a monkey is not an author, for purposes of the Constitution, and that you cannot have copyrights owned by a monkey. So this is kind of an unusual twist. And now, you know, whether a robot or a computer can be entitled to something completely created by a machine, can that be entitled to copyright protection? And the answer is no.
02:07
And, and so some of these things, though, what I'm reading is, for some of them, it is a combination. So in other words, there will be an author who actually generates the content, but the AI is or the—whatever it is, the robot—is creating the artwork, and vice versa. So how does that work?
Yeah. And you bring up an interesting point. And that's what, that's what the law is grappling with. You know, for example, is the computer a tool? And when the computer becomes a tool, is it functioning, kind of like a paintbrush? Like, nobody would deny a painter a copyright to a painting, because he used a paintbrush. That would seem absurd. There's a Supreme Court case, way back in 1884, that held that photographers could get copyright protection for their, for their photos for pictures. And clearly a photographer uses, you know, a tool, a camera. And of course, as cameras get more and more sophisticated, the human input becomes less and less. I mean, they now have, you know, clearly they can get rid of red eyes, they can change shading. The amount of skill needed by the human to operate goes down. And that's where this recent case is Zarya of the Dawn, the author used AI to generate art. And the difference is that, you know, in that case, the Copyright Office held that it's not, there's no copyright protection for that art, because the human interaction was extremely limited. And there was no control. Like, unlike a painter, you might use a paintbrush, but the painter is in control of the final product, the brush is just a tool. And the same thing with a camera. Yes, the camera is a tool that takes pictures. But there's a lot of control the photographer has. He's holding the camera, he is now, you know, aiming it, directing it, choosing the time, working with the lighting. So there's a lot of ongoing—the copyright office uses the term iterative control that the author has. Now the less control an author has, and the more the tool or the machine or the robot, the more that machine does, and the less the human does, that's when it becomes problematic. And in this case, they use this AI tool called Midjourney that creates images. But there's no way for an author to edit the image after it's created. There's no input, there's really you're stuck with the results that are delivered. And in that case, it was a clear case where the computer was doing all of the so-called creative work, there's really very little on the part of the human. And in that case, it was denied. There's no copyright protection for computer generated artwork.
Well, my question was that one would be, and this I believe was a cartoon or comic I guess. Anyway, my question with that one would be, doesn't the AI or and or robot have to have a description of some sort generated by the writer that describes said person in comics? And so therefore,isn't that sort of murky there? Do you see what I'm saying?
Yeah, so then I know exactly what you're saying. So the creativity is in the input, right. So that's what the author's argument is that, hey, listen, choosing what terms to put in is where the creativity comes in. And unfortunately, because this particular app, Midjourney, specifically has a user's guide that says it doesn't understand words, grammar, or intent the same way as a human does. The results are generated more or less randomly based on the input, and they didn't feel that input had enough control over the output. It may be different. In the case of—you mentioned chatGPT. So with chatGPT, a big difference is when you get the final, and just for the benefit of some of your listeners, chatGPT is a new tool that basically allows the computer to write sentences, content, in response to a natural language text. But the difference is, when you get that content written, the author is permitted to edit it, revise it, make changes. So it's really used—you can use it as a first draft, and then put some human creativity into the final product. And it doesn't take a lot—the creativity to get a copyright is low, but it's not zero. And in the case of this Zarya of the Dawn, it seemed as if the human creativity was really very close to zero. And there is no, the big distinction might be there's no way using that application, Midjourney, there's no way to edit the images once they're presented. I think the case would be different if the author took the images that were there, used Photoshop or some other program to make changes to it. Even if they're minor changes, then they've put in the human creativity needed even if all you're doing is revising the work created by a computer. But if you take the work created entirely by the computer, you don't revise it, you don't edit it, you don't make any changes, then that's different than somebody using chatGPT to write an essay for them, but then they go back and add headings and revise the text and add an intro or, or somehow involve themselves as a as a human in the creativity. That's where chatGPT is different than this particular case regarding Mid journey.
Thanks for explaining that. But now, John, aren’t we just on the tip of the iceberg when it comes to copyright and patents for artificial intelligence generated art, literature and that sort of thing? I mean, do you foresee some laws changing as a result?
08:51
Oh, absolutely. Like anytime there's a new technology there's this huge lag before, before laws change. And it's not even necessarily laws. For example, Amazon, currently the largest bookseller in the world, does not have any restrictions on books written via chatGPT. There is no requirement even that an author reveal that chatGPT has been used to write a book. My prediction is that that's going to change because ultimately consumers are going to demand to know which books are written by an author versus which are generated exclusively via chatGPT. So, the laws change. I mean, when drones first came out, there were no restrictions on where you could fly a drone. There were people, the early adaptors, that were flying drones in state parks and getting unbelievable coverage at, you know, national state parks… And then finally the laws caught on and it's like, hey, wait, we can't—and now you're you're not permitted to do that. So anytime you have a new technology there is this lag and I think we’re seeing that with copyright law right now.
10:04
So I write something, or I have the computer, write an essay for me, and I hand it in. Will there be the capabilities kind of like Shazam was when somebody can enter that and it'll look out about the vast world and go, nope, that's already been written, it was written by Darrell over in blanket. You cheated. Is that going to be possible?
Yeah. Oh, it's huge, you know, the implications to education are incredible, like, right, like plagiarism? How do teachers find out? There's now new apps being developed to help identify when chatGPT has been used. But you know, that that's going to be a completely game changing app in terms of teaching writing and analytical thinking. I think it's going to be a huge loss in terms of people developing cognitive skills. I mean, that's sometimes just the nature of what technology does. Like I have, I have teenagers that have trouble reversing their car if the rear view camera is not working, right? We grew up without rear view cameras, so we've developed that skill. If the rear view camera doesn't work, we're still able to function. If the some of the safety features, they're great, as long as they work. But if you don't have sideview—I guess, so that, if you're changing lanes, like you know the way the blind spot detectors, if you don't have that in your vehicle, and you've never developed the skill to look in both mirrors, then you're in trouble. And that's, we're gonna see that with the advent of the smarter computers get, and the more sophisticated chatGPT gets, it's just, it's the writing skills and cognitive thinking and analytics is going to suffer as a result. And there's, you know, unless there's some way for educators and teachers to get around that. I mean, I possibly even see a lot more less writing assignments, where the teacher can make sure the professor can make sure that students are not using chatGPT.
You're gonna have to write it in hand with your own handwriting.
12:28
Yeah, you know, all that's old is new again. You might see, you know, something going back to handwritten essays again. So a lot of schools stopped teaching, script and cursive. But who knows in terms of how the educational system is going to adapt to preventing cheating and making sure that students really learn those analytical skills, that might be one way where handwritten essays make a comeback, at least in the educational environment.
I was gonna ask you this, you know, you're a professor, have you already experienced students using chatGPT?
13:11
Yeah, I haven't. Because again, for law school, the exams are always in person. No computers, they use blue books, handwritten examinations. It has impacted a lot of professors, including myself. I, you know, I would hesitate now in the, at home assignments, right, because that's something where I have no control over whether someone is using chatGPT. At this point, it's not sophisticated enough that a student can rely entirely on it. But I just liked the level playing field of having a pen and a booklet and a closed environment room, and everyone's got the same opportunity.
Agreed. Well, thank you so much. It has been an interesting conversation. We've been speaking this morning to John Rizvi. He is known as The Patent Professor. He's an adjunct professor of patent law at Nova Southeastern Law School in Florida. And John, is there a place our listeners may go to find you online?
Yes, The Patent Professor is our website, but Instagram, Facebook, anywhere on all social media our handle is The Patent Professor.
All right, John Rizvi, thank you for taking the time.
14:41
Thank you.