If an AI takes a picture, who owns the copyright?

[Snowboarder_Bo]

@TimeWinder

Not at all. By that logic ALL photographs would be in the public domain, because the human didn’t open and close the shutter themselves, they allowed a mechanical system with unknown and variable response time to do it for him/her.

Not accurate, since the operator has presumably set the response time, shutter speed, etc. Even not changing them is a decision that a person can make. And it presumes a degree of removal that can’t be reasonably argued, IMO.

@TimeWinder

A human bought the camera intending to use it for this specific purpose. The human chose the camera’s position and framing, decided when to turn it on and off, chose the events that would take place in it’s field of view.

Not all of this is accurate either, IMO. If the AI can zoom in, it is choosing the frame. The human may or may not be able to argue that they “chose the events that would take place in it’s field of view”; I suspect that most people will not be actively directing people to step into the camera’s field every moment it’s on. I mentioned, for example, setting it up during a party and letting the AI take pictures.

@TimeWinder

None of these things are anything like the monkey case whatsoever, where an unexpected sentient being took unanticipated actions and happened to produce something useful; there was no human intent involved at all.

So David Slater didn’t bring a camera to Indonesia to take pictures of wildlife? He didn’t choose the film stock, the lens or the aperture setting? In fact, he did more selecting and had more input on the photo than many people would have with an AI, yet his claim was denied.

@TimeWinder

As others have pointed out, there are tons of technologies now that separate the action of the button press from the shutter release (sometimes by vast amounts of time), and nobody’s claiming that those photos aren’t copyrighted. A hundred non-human-controlled actions happen every time you take a photo.

And the human is presumed to have had the ability to dictate how those functions should operate and thus have controlled them, even if they decided that the current setting was the one they wanted and thus made no active changes.

In the AI case, tho, the AI will decide the aperture, the focus, the zoom, the frame and the subject. It will only not select it’s placement or when it is on (since we don’t make devices that can enable themselves (yet)).

As I say tho, the AI is having more input into the image capture than the human is.

@TimeWinder

Mind you, I think the monkey case was wrongly decided, but either way it has nothing to do with these auto cameras. A more analogous situation might be automated security camera footage, and I know I’ve seen copyright marks on some of those.

Well on that we agree, although I’m not certain we will agree on why it was wrong and how.

And the thing about the security camera footage is that there is no AI directing the variables in the shot.

The rule on copyright claims is that

Similarly, the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author.

From the MSCD I am not at all certain that placing a camera and turning on the AI is enough “creative input” or “intervention from a human author”. And if there is sufficient intervention, I think it more likely that the author of the software has a higher claim to that than the person who set it down and turned it on. Obviously, EULA and TOS would come into play, but let’s face it: most people don’t read them and don’t much care about them, being pretty assured that no one is going to knock their door down and arrest them for some violation of such.

And lots of people simply don’t care. Facebook has a huge image library available to them that they acquired for free, simply by telling people they could post photos.

[Snowboarder_Bo]

nm

[Darren_Garrison]

@Snowboarder_Bo

In the AI case, tho, the AI will decide the aperture, the focus, the zoom, the frame and the subject. It will only not select it’s placement or when it is on (since we don’t make devices that can enable themselves (yet)).

Are you talking about this Google product specifically, or are you talking about some hypothetical theoretical future AI that does not yet exist? Because the Google product does not choose zoom–it has no zoom. Therefore, it does not choose frame–the “frame” is where it is pointing. It also arguably does not choose subject–“subject” is who walks in front of it. And in modern cameras, algorithms already choose the aperture, the focus, the ISO, and the shutter speed unless you specifically switch it into partial or full manual mode (and cheap cameras don’t even have that option.)

I’ll give this conceit–if you are talking about some imagined future AI that can pass a Turing test that uses as much volition and choice as you are asserting, then I’ll stand up for the damn thing getting the copyright (and civil rights.)

[AI_Proofreader]

@Darren_Garrison

But you *are *pressing a button that says “take pictures when you see a smiling face.” He has yet to prove how this is materially different than current camera options that include “take a photo in 10 seconds”, “take a photo every 5 minutes”, and “take a photo when you sense motion.”

Agreed… setting the feature constitutes the *intent to take a photo. The fact that an algorithm/monkey/gust of wind/Act of God was then the instrument for the machine carrying out that intent, doesn’t matter. You set the machine to produce a photo, it produced it. Clearly you’re the person the photo belongs to.

If a phone had no such feature available and *chose to take pictures on its own with no original input from a human… then the photographer is the phone. Also you need to kill it with fire, because robot apocalypse.

[Mosier]

I think the main difference between the monkey and the AI is that the human is taking deliberate action by placing and activating the camera. In the monkey thief case, the pictures were entirely accidental, not at all caused by anything the human intended or took any deliberate action to bring about.

If you still think the human would have no legal right to the photos taken by the A.I., could you explain how this scenario differs fundamentally from motion tracking security cameras? Why would the program designed to capture specific moments in your scenario prevent the owner from claiming copyright, but the program designed to capture specific moments in my scenario (the motion-sensitive camera) wouldn’t?

[RaftPeople]

@Mosier

I think the main difference between the monkey and the AI is that the human is taking deliberate action by placing and activating the camera. In the monkey thief case, the pictures were entirely accidental, not at all caused by anything the human intended or took any deliberate action to bring about.

The monkey case was entirely accidental? I thought the photographer flew to a remote location, hiked into the jungle and setup all of the camera equipment so that the monkeys would have the opportunity to press the button, seems like a pretty similar situation.

[Mosier]

@RaftPeople

The monkey case was entirely accidental? I thought the photographer flew to a remote location, hiked into the jungle and setup all of the camera equipment so that the monkeys would have the opportunity to press the button, seems like a pretty similar situation.

Interesting! I didn’t know that was the case. I thought the monkey just stole his camera and ran off with it.

[Snowboarder_Bo]

@Snowboarder_Bo

It is settled: copyright can only beheld by a legal person. AI and non-humans are, at present, not legal persons in any nation that I know of.Aye; we can make (or unmake) laws as needed and someday there likely will be laws that address this.

It seems I had wildly incomplete information:

Aiva Technologies is one of the leading startups in the field of AI music composition. It was founded just last year in Luxembourg and London by Pierre Barreau, Denis Shtefan, Arnaud Decker, and Vincent Barreau.

They have created an AI called “Aiva” (Artificial Intelligence Virtual Artist) and taught it how to compose classical music – an emotional art which is usually considered to be a uniquely human quality. Aiva’s musical pieces are used as soundtracks for film directors, advertising agencies, and even game studios.

Having already released its first album called Genesis, as well as many single tracks, Aiva recently became the first AI ever to officially acquire the worldwide status of Composer. It was registered under the France and Luxembourg authors’ right society (SACEM), where all of its works reside with a copyright to its own name.

I’m not really clear on how that works, tho, or how’s different from a band’s work being registered under the band’s name rather than under the names of the individual’s who are members of the band.

[Darren_Garrison]

Just saw a commercial for this.

[Kadmos1]

Not the estate of the creator of Tarzan.

[running_coach]

@Kadmos1

Not the estate of the creator of Tarzan.

What’s that have to do with the topic?

[Snowboarder_Bo]

@Kadmos1

Not the estate of the creator of Tarzan.

I agree that if an AI takes a picture, whatever estate is left of Edgar Rice Burroughs’ should not control the copyright.

[Alley_Dweller]

@Snowboarder_Bo

This is how I follow the legal logic chain, too. In which case, Google (or anyone who can get ahold of the image) can use it in any way they see fit. How many years until Getty goes out of business if that’s the case?

Getty makes sure it has the necessary model releases on file for all of its photographs (or it labels them as “for editorial use only”). As I’m sure you know, model releases have nothing to do with copyright.

Yes, you do not need model releases for all uses, but they can’t just use all photos “in any way they see fit.” Even if they have a shrink-wrap license agreement or the purchaser has to click “yes I agree” when installing the software, the purchaser can’t give away the rights of other people captured in the photograph. Google would need to have some way to control that only the persons who agreed to the terms appeared in photographs they used for product endorsements, etc.

[Kadmos1]

running coach, a mid-March 2017 thread called “Can you please critique this letter to the estate of Tarzan’s creator? Long posts before the letter.” that was started by me: boards.straightdope.com/sdmb/showthread.php?t=821892.

[Bone]

@Kadmos1

Not the estate of the creator of Tarzan.

@Kadmos1

running coach, a mid-March 2017 thread called “Can you please critique this letter to the estate of Tarzan’s creator? Long posts before the letter.” that was started by me: boards.straightdope.com/sdmb/showthread.php?t=821892.

Drop this hijack about Tarzan. Don’t spam the boards about Tarzan.

[/moderating]

[Darren_Garrison]

@Snowboarder_Bo

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.

Update: no they don’t.