Google has a new product called the Google Clips camera which is unique in that you don’t have to take the pictures, the camera will take them all by itself.
I can, however, also see where courts could rule that Google owns the copyright on any images captured by their AI, since they created the software and directed it on how and when to take pictures.
Certainly it seems that the person who owns the camera and turns it on has no more claim to creating the photo than David Slater had.
“How much individual human direction is necessary for a work to be copyrightable?” may be the crux of the biscuit here.
The person who owns the camera? Having a camera programmed to take photos in x condition doesn’t sound much different than having a camera programmed to take photos every x seconds, or when motion is sensed. This is IMHO a major, major non-issue.
Okay, so let’s take this prolly-not-too-far-off hypothetical: I buy a Google Clip camera. I place it on a shelf and turn it on to use the AI to take the pictures. The images are uploaded to whatever cloud service Google provides with the camera. Two months later, I see one of those pictures in an advertisement online. I say the picture is mine, because I own the camera and I turned it on. Google says the image is theirs, since they wrote the software and designed the hardware that captured the picture.
How is my involvement any more intrinsic to the creation of the images than David Slater’s involvement in the creation of the monkey selfie?
I really don’t see the issue. Many (if not most?) modern cameras already have software for facial recognition–Google is simply hyping actually taking a photo after recognizing a face instead of just setting the focus, shutter speed, etc. Nobody is going to claim owning a photo because they wrote the firmware that operates the camera.
Google might argue that they provided the artistic expression of each photo by making algorithms that looked only for certain situations and conditions before taking a picture. The owner might argue the same because of camera placement, lighting, and using whatever settings the camera allows. Might go either way.
Other than the monkey selfie, David Bowie provides another case. He’s been known to cut up pieces of paper with writing on them, then pick random words from a hat for some of his lyrics. Before his death last year, he began using a computerized version of the process. Are those lyrics copyrightable?
That depends on how much artistic control he exerted, how much “power of creation” he used. Did he accept some of the generated lyrics while rejecting others? Did he decide which words would go into the hat beforehand? Was randomness, itself, part of what he was trying to convey?
The issue is that the arguments that you’re putting forth were struck down by a court as insufficient for the purpose of copyright. David Slater clearly owned the camera in question. He selected the film stock and all the setting on the camera. He even purposely went to Indonesia with his camera, meaning to take pictures. The only things he did not direct were a)what the camera was pointed at and b)when the image was captured.
And because of those things, the court ruled that he had no claim to the copyright.
In my hypothetical, the person using the camera has had arguably even less control over the image being captured; why should they retain a copyright for the images captured on their camera?
Some camera apps have speech recognition so that if you say a specific word (such as “cheese”) a photo is taken. Does this mean that the guy who wrote that voice recognition algorithm has a claim to the copyright to any photos that you take? Hopefully, you will admit how silly that sounds. Well, just because Google uses a heavy-handed buzz term like “AI” to describe their algorithm doesn’t make it any less silly.
No, it absolutely did not happen that the programmer of camera firmware was granted copyright ownership of photos taken using that camera. It does not happen for autofocus firmware, it does not happen for light metering firmware, it does not happen for shutter speed firmware, it does not happen for red-eye reduction firmware, and it will not happen for facial recognition firmware.
Wow. I have no idea what you’re so angry about, but I’m reasonably certain it isn’t me, David Slater or the monkey.
Your posts in this thread have been characterized not just by that anger, but also by a distinct lack of understanding of the issues at play in copyright law and in the MSCD and a lack of or unwillingness to generalize enough to see the commonalities in my hypothetical and the actual MSCD.
I have been trying to explain those issues, but you seem unwilling to accept the proceedings of the MSCD and the court’s decision as facts.
I linked to the Wikipedia page, thinking there was sufficient information there for people to enter this discussion at least reasonably prepared, but that seems to have been erroneous on my part.
I’ll try again: the parallels between an AI taking a picture and a monkey taking a picture are, or should be, obvious: it is not a human deciding when to capture the image.
Current law says that non-humans cannot claim copyright on things they have created. The MSCD result said that in this case, the human did not have a claim on the image created by a non-human, even tho the human’s actions made the image capture possible, because it was not sufficiently directed by said human because it was the monkey who pressed the shutter button. The fact that the film stock, aperture settings, type of lens, etc. were all selected by David Slater were deemed not important enough for him to claim copyright over the image.
Google’s Clips camera will set the aperture, exposure time, shutter speed, zoom, etc. The AI will decide when and what images will be captured. A person taking pictures with one on full AI-auto is arguably doing less work and having less input than Mr. Slater did with his camera.
So who gets to claim that copyright?
It is important, IMO, because many people will not think about this until after their family birthday party pictures or BBQ or wedding photos end up in advertisements because their pictures will either have no copyright or the copyright will be assigned to Google, since they had more input and direction, including when and what image would be captured, than the owner of the camera.
(And note that you may not actually be buying the camera, nowadays you might very well simply be paying a fee for the use of the software and hardware without any actual ownership of either. This is a whole separate issue that may complicate things…)
There is no “anger” here, that is pure projection on your part. I do, though, think that you have an extremely tenuous grasp on what the “AI” in this camera represents–it is a small, incremental change from various sorts of automation found in all modern cameras–it is a small qualitative change, not a quantitative one. It is marketing hype. It is not HAL 9000.
But let’s not let this be simply a two-way discussion (that I already made a liar of myself over by saying that I would bow out on.) Anybody else think that a camera manufacturer will claim copyright over photos taken with a camera because their camera has facial recognition firmware that can be programmed to trigger image captures?
Yeah, on a camera that’s going to take pictures of people in an area where they might forget it’s on, that’s pretty much a necessity if Google doesn’t want a lawsuit the first time somebody produces inadvertent amateur voyeur porn.
I think there is a very arguable difference. The photographer didn’t set up the situation where the monkey would take the picture. The camera was taken from him without his permission. A lot of the arguments I saw on the copyright issue mentioned this as a reason he could not claim copyright over the image.
I also note that whether the monkey himself owned the copyright was settled out of court, with a contribution to some wild life fund. So it’s not settled it didn’t own the copyright, just that the man didn’t. If the AI can hold the copyright, then there is no reason there can’t be licensing agreements that hand that copyright over to the user or to Google.
Then there’s also copyright on one’s likeness, which you can register. It has been used for “revenge porn” situations. The copyright could then have to do with the subject, absent any other copyright. You don’t have to have the whole work copyrighted: note that no one has the copyright on the It’s a Wonderful Life movie, just the story it was adapted from. But that’s enough to control the ability to copy the movie.
And, finally, there’s always the possibility of introducing new law, whether by the courts (which do make law in a common law system) or by the legislatures. This isn’t a constitutional issue, so the new law would control. And since AI will be doing more and more things, it’s rather necessary that it be addressed in law.
Yeah, I thought the monkey picture wasn’t so much about who took the picture, but who set up the shot. There are plenty of motion-detection photos where the camera owner retains copyright, because they set up the shot, and the deer or raccoon or whatever just triggered the shutter.
With the monkey photo, the monkey actually took the shot. He framed the picture, smiled, and clicked the shutter. That’s photography, and David Slater had nothing to do with it. If setting the aperture and other settings was photography, most camera owners in history wouldn’t retain copyright on their own family photos, because most people don’t fiddle with any settings at all.
A more relevant hypothetical device, in my opinion, would be something like the Google camera from the OP, but mounted on a drone which flew around and found the best angles to take pictures from. In that case, the camera owner isn’t the one setting up the shot, which will just be triggered automatically at a later time as in the case of a motion-detecting camera. The “AI” itself would be pointing the camera and making the decisions about the best camera angles, background, and lighting, in addition to when the shutter operates.
And I would argue that in the case of a robot photographer, nobody retains copyright, it certainly shouldn’t revert to Google. And people still retain a right to privacy, so they shoudn’t be surprised by their private photos showing up in advertisements.
I. getting an angry vibe from the emphatic and dogged way that you’re arguing against something that hasn’t been argued for. And I think now that you’re focusing on specifics without understanding that the OP is about the bigger picture.
“Anybody else”? “Else”? Who has taken that position that there would be “anyone else”? :dubious:
Correct, also AIUI. The fact that a person sets up a camera with a trigger to be activated is enough to say that the person is controlling the image, even if the trigger is a random event. Mr. Slater’s problem was that he admitted that he did not intend for the monkey to take his camera. Presumably if he had said something like (preferably on a video recording) “Goddammit monkey; now that you have my camera, take some ddamned pictures!” that would have been enough for the court to say “okay, he intended for the monkey to snap photos” and he would have been granted copyright.
And I think that most camera owners would retain copyright in that last bit because they deliberately left the settings in place.
The thing is, AI doesn’t own itself. It is an agent of someone (or some company): whoever created it. With the monkey selfie, no one owns the monkey, so there was no external party trying to claim ownership of the monkey’s creation. In the case of AI, why wouldn’t the creator of the AI have a claim on copyright of the AI’s creations?