Copyright in AI-assisted works
According to an article on Arstechnica, the US Copyright Office has granted its first ever copyright registration for an AI-assisted artwork.
This may seem like a momentous step, but it is important to remember a few critical points:
- The copyright registration is for an entire graphic novel, not just for a single image. The novel as a whole involved significant human input in relation to the story, layout and arrangement of the images.
- The applicant Kris Kashtanova made it clear that the artwork in the novel was AI-assisted, not created entirely by the AI.
- The registration credits the individual Ms Kashtanova rather than the AI as the author. This is the key difference from the situation in the news in June 2022 when Stephen Thaler tried to register copyright with the AI as the author. In that case, the US Copyright Office refused registration on the basis that a machine could not be the author of a copyright work.
Australian position on AI-assisted works
In Australia, copyright can only arise under the Copyright Act 1968 if the author is a “qualified person”. That is, “an Australian citizen or a person resident in Australia”, with additional protection along similar lines being afforded to non-Australian works.
“Author” is not defined in legislation, other than that “in relation to a photograph, [author] means the person who took the photograph. That definition specific to photographers was merely to change the pre-1 May 1969 position when the author of a photograph was the owner of the material on which the photo was taken.
AI assistance is not new
When we step back from this latest scenario, we know that machine involvement in the creation of copyright works is not new. In the world of photography, we often leave the setting of one or more of aperture, shutter speed, ISO and exposure to the camera. Those settings have significant impact on the appearance of the photo and yet often sit outside the author’s control.
After the point of image capture, we might make more adjustments to the image. Software (eg Adobe Photoshop) might be used to create content aware fill to cover the gap created by removing unwanted elements from the photo. The software doesn’t know what was actually behind the person that is removed from the scene, so it takes a guess and makes it up.
While there might be some creators who look down their noses at those who don’t photograph in full manual mode or those who perceive post-processing as cheating, these aids are not controversial from a copyright perspective. And frankly, they pre-date the digital era by decades, with Josef Stalin famously being a fan of significant photo editing.
Limits to acceptable assistance
Of course, there are limits to when a person can still claim authorship of something that they did not fully create themselves. That is not unique to AI involvement.
In 2014, the US Copyright Office determined that a photo was taken by a monkey rather than by the photographer David Slater who had set up the camera. Accordingly, the Copyright Office decided that neither Mr Slater nor the monkey should hold copyright. The Compendium of US Copyright Office Practices in Chapter 300 states that it “will not register works produced by nature, animals, or plants.” Further:
“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. The crucial question is “whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.” (Chapter 313.2)
This then becomes a question of degree. The UK Intellectual Property Office expressed it as:
“The question as to whether the photographer owns copyright is more complex. It depends on whether the photographer has made a creative contribution to the work and this is a decision which must be made by the courts.”
This is also likely to be the position in Australia, given that we come back to the same question of whether the author was a qualified person.
The take-away
Herein lies the rub. In the case of Ms Kashtanova’s graphic novel, she had made significant decisions about that work. In the case of Mr Slater’s monkey photo, the US Copyright Office determined that he had not done so.
When you encounter an AI-generated or AI-assisted work, it may be almost impossible to determine whether copyright subsists unless you know more about the creative process. Stock image libraries such as Getty have already announced that they will cease licensing AI-generated images for this reason.
If you’re a budding AI user and enjoy seeing what the AI can create for you, maybe consider whether you own copyright in “your” creation before you share it online. You might not be able to “get the genie back in the bottle” if you don’t own copyright.
Conversely, if you want to use an existing AI-generated work, there is considerable risk in assuming that no copyright subsists and that it is therefore free to use.
If you’re after more specific advice on the question of copyright ownership, reach out to the Macpherson Kelley IP team.
The information contained in this article is general in nature and cannot be relied on as legal advice nor does it create an engagement. Please contact one of our lawyers listed above for advice about your specific situation.
stay up to date with our news & insights
Copyright in AI-assisted works
According to an article on Arstechnica, the US Copyright Office has granted its first ever copyright registration for an AI-assisted artwork.
This may seem like a momentous step, but it is important to remember a few critical points:
- The copyright registration is for an entire graphic novel, not just for a single image. The novel as a whole involved significant human input in relation to the story, layout and arrangement of the images.
- The applicant Kris Kashtanova made it clear that the artwork in the novel was AI-assisted, not created entirely by the AI.
- The registration credits the individual Ms Kashtanova rather than the AI as the author. This is the key difference from the situation in the news in June 2022 when Stephen Thaler tried to register copyright with the AI as the author. In that case, the US Copyright Office refused registration on the basis that a machine could not be the author of a copyright work.
Australian position on AI-assisted works
In Australia, copyright can only arise under the Copyright Act 1968 if the author is a “qualified person”. That is, “an Australian citizen or a person resident in Australia”, with additional protection along similar lines being afforded to non-Australian works.
“Author” is not defined in legislation, other than that “in relation to a photograph, [author] means the person who took the photograph. That definition specific to photographers was merely to change the pre-1 May 1969 position when the author of a photograph was the owner of the material on which the photo was taken.
AI assistance is not new
When we step back from this latest scenario, we know that machine involvement in the creation of copyright works is not new. In the world of photography, we often leave the setting of one or more of aperture, shutter speed, ISO and exposure to the camera. Those settings have significant impact on the appearance of the photo and yet often sit outside the author’s control.
After the point of image capture, we might make more adjustments to the image. Software (eg Adobe Photoshop) might be used to create content aware fill to cover the gap created by removing unwanted elements from the photo. The software doesn’t know what was actually behind the person that is removed from the scene, so it takes a guess and makes it up.
While there might be some creators who look down their noses at those who don’t photograph in full manual mode or those who perceive post-processing as cheating, these aids are not controversial from a copyright perspective. And frankly, they pre-date the digital era by decades, with Josef Stalin famously being a fan of significant photo editing.
Limits to acceptable assistance
Of course, there are limits to when a person can still claim authorship of something that they did not fully create themselves. That is not unique to AI involvement.
In 2014, the US Copyright Office determined that a photo was taken by a monkey rather than by the photographer David Slater who had set up the camera. Accordingly, the Copyright Office decided that neither Mr Slater nor the monkey should hold copyright. The Compendium of US Copyright Office Practices in Chapter 300 states that it “will not register works produced by nature, animals, or plants.” Further:
“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. The crucial question is “whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.” (Chapter 313.2)
This then becomes a question of degree. The UK Intellectual Property Office expressed it as:
“The question as to whether the photographer owns copyright is more complex. It depends on whether the photographer has made a creative contribution to the work and this is a decision which must be made by the courts.”
This is also likely to be the position in Australia, given that we come back to the same question of whether the author was a qualified person.
The take-away
Herein lies the rub. In the case of Ms Kashtanova’s graphic novel, she had made significant decisions about that work. In the case of Mr Slater’s monkey photo, the US Copyright Office determined that he had not done so.
When you encounter an AI-generated or AI-assisted work, it may be almost impossible to determine whether copyright subsists unless you know more about the creative process. Stock image libraries such as Getty have already announced that they will cease licensing AI-generated images for this reason.
If you’re a budding AI user and enjoy seeing what the AI can create for you, maybe consider whether you own copyright in “your” creation before you share it online. You might not be able to “get the genie back in the bottle” if you don’t own copyright.
Conversely, if you want to use an existing AI-generated work, there is considerable risk in assuming that no copyright subsists and that it is therefore free to use.
If you’re after more specific advice on the question of copyright ownership, reach out to the Macpherson Kelley IP team.