What if your Artificial Intelligence creates another AI? Who owns it?

Artificial Intelligence, its everywhere. Despite initial expectations of slow progress, recent updates in the area of AI’s creative capabilities reveal a variety of original works created by intelligent algorithms. Some algorithms have been known to create “journalistic content” and others have achieved what may be described as “artistic expression”. Latest developments go even further and suggest that “creative writing” is also not reserved for humans anymore and a proof of that is “Sunspring” – the first ever short film that uses a script written entirely by an algorithm.

So, who should be the author of these inventions? You? The AI? Or someone else?

The Human

Giving rights to one of the humans involved in the process of creation – the programmer or the user of an AI – is the most expected and the least controversial solution.

The biggest argument in favor of this approach is that the programmer is the one who is closest to the creative process and understands the process behind the AI invention. This also sets very well with the main idea behind the copyright as well – by granting authorship, you can motivate the creative development and production.

Now, what if the decisions of the AI cannot be explained? For example, the developers that are working on the YouTube’s algorithm have no idea what or why the changes are made in the algorithm. So, this approach ignores the fact that the programmers have limited control and little to no understanding of the creative process.
There are lots of other things, like the programmers getting rewarded twice - once for creating the AI and then for its outputs, that make this a bad option.

The user

Now, what if we granted the authorship to the user of the AI? The idea here is that the user who created an original work by using the AI should be rewarded the authorship of the invention.

But the problem here is that there are a ton of AI applications in the market that can generate an original work with just a click of a button. So, it is really hard to justify that the creation consists of an original work when all it took was a couple of clicks.

The machine

Another option in front of us is to grant the authorship to the machine. This is justifiable because, obviously, the AI has done most of the work and it should be granted the authorship of its work.

The biggest argument against this option is that machines cannot exercise their rights as an owner. This is problematic because an algorithm needs personhood in the eyes of the law to be legally deemed as an “author”.
Another requirement for the authorship is that the invention needs to have an “imprint of the author’s personality”.

This is a huge obstacle because the machine would have to possess “human-like-personality” in order to be considered as an author. Also, the main goal of the Intellectual Property law is to provide incentives for authors to create more and thus expand creative expressions and benefit society. But, machines do not need incentives to create.

Joint Authorship

The idea here is to grant the authorship to both the programmer and the machine. The machine can get the title of the author whereas the programmer can bear the fruits of the authorship.

But the problem here is that this in turn opens the gate for other relevant stakeholders, i.e. the provider of the hardware or the trainer of the AI or the user. Finally, joint authorship, usually requires the contributions of the authors to be distinguishable, which in the cases of complex computational techniques such as machine learning would be close to impossible.

No author

Leaving the creation of a creative AI in the Public Domain is an easy and simple thing to do and follows the natural logic of copyrights.

The biggest benefit of this approach is that the individuals will use AI creations without fearing copyright infringements. But, the problem here is that this contradicts the main goal of the copyright laws – to create benefits for society. So, if the AI inventions are left in the Public domain, it will diminish the incentives to invest and develop AI technologies, ultimately leaving society worse off.

The biggest argument against this approach is that exporting the output of creative AIs in the public realm would deprive many stakeholders of the reward they expect and deserve. After all, developing and deploying a form of creative AI is a long, time-consuming and financially expensive process and all parties involved are entitled to economical gains in order to benefit from the whole project.

There are a lot of issues when granting authorship to an invention of an Artificial Intelligence. And, these issues needs to be addressed as quickly as possible because Artificial Intelligence's are creating original works nowadays. Our Intellectual Property laws and Patent laws have to get updated.