Can Autonomously-Generated Content be Copyrighted, Redux
Back in 2013 I was one of the first to try to answer the question of whether works generated by AIs were covered by copyright law. The answer back then was "it depends", and absent definitive court rulings we’re still waiting to see how this issue will play out.
In the years since we’ve seen the monkey selfie story reinforce the idea that only a person can author a work, so one might think this matter settled. However, that case involved an animal, and there is existing case law and copyright law that covers databases, so this is very much up in the air.
Last month the 1709 blog analyzed this issue from the perspective of EU law, and expressed doubt that an AI work can be copyrighted.
Another argument against the proposition that also non-human authors can be regarded as potentially eligible for copyright protection in their works is that, at the EU level, the general rule is that concepts used in different directives must in principle have the same meaning. So, in Football Association Premier League, C-403/08 and C-429/08, the CJEU clarified that if a directive in based on rules and principles already laid down in other directives, “given the requirements of unity of the European Union legal order and its coherence, the concepts used by that body of directives must have the same meaning, unless the European Union legislature has, in a specific legislative context, expressed a different intention.”
In the case of the InfoSoc Directive (as well as the Term Directive, with specific regard to photographs) this might mean that, similarly to what is instead expressly stated Software Directive, the concept of author implies that this is a human being. This is because, as the CJEU clarified in Infopaq, C-5/08, the InfoSoc Directive, as well as the Term Directive, is based – inter alia – on the same principles and rules laid down in the Software Directive.
It is impossible to definitively answer this question until after it has been hashed out in the courts, but at this point there is nothing to suggest that an AU-authored work could be copyrighted.
One could point to existing case law on copyrights on databases, but the existing law (in the UK) and case law (in the US) cover situations where a person put data into a database and then claimed a copyright on the database.
This means that on a fundamental level, existing case law is not strictly relevant to the issue of AI-created works where no human was involved in authorship.
image by michaelmusashi
vrf January 4, 2018 um 1:37 pm
I wonder if it would be any different from procedural software used to generate artwork. The software still relies on parameters, rules and ranges established by a human designer. A robot never really creates anything on its own.
Nate Hoffelder January 4, 2018 um 4:11 pm
"A robot never really creates anything on its own."
The term robot is so broad that I don’t think you can say that with certainty. Yes, it is true when applied to some bots like Twitter bots, but don’t neural networks operate independently, and grow into new structures?
Their output might count as being created without human intervention.
vrf January 4, 2018 um 5:05 pm
Neural networks really can’t do much without human intervention–output is always derived or influenced by input–but I agree that copyright lawyers and judges may see things differently. Surely having a robot follow some bland rules about grammar and language construction isn’t in the same creative league as an author or painter.
Tom Wood January 4, 2018 um 7:14 pm
Saudi Arabia’s (ironic) granting of citizenship status to a robot aside, robots and artificial intelligence programs are property. If said property creates something new, it would most likely fall under 'work for hire' and the copyright would be held by the owner of the property.