Triple-A — Art woes and Artificial intelligence Authorship
I was going for a ‘Rise of the Planet of Apes but make it every sci-fi dystopian movie with AI taking over’ as the vibe for this article. Drama aside, you would be surprised to see how far Artificial Intelligence (AI hence) has come IRL. When I think of AI, I think of JARVIS or EDITH or Ultron (Marvel Nerd alert) but there are plenty of sci-fi movies that have artificial intelligence as one of the antagonists or main cast members (ala Matrix, I-Robot, Blade Runner, Ex-Machina, Westworld, Her, Tron, and the list goes on). Despite the popularity in pop culture and the whole idea that AI will take over the world, in the current time, we have not advanced to the stage where AI replaces humans (AI superintelligence) or show human-like intelligence (AI General intelligence). You could read more on the classification here.
The AI applications used by us in the present time such as Amazon’s Alexa, Apple’s Siri, Google Translate, Google Assistant are classified as ‘narrow’ and are ones that merely “pull information from a specific data set” and “do not perform a task beyond the one designed to perform”. In gist, they cannot think for themselves and are far away from human-like intelligence. What complicates matters however is what if you taught the AI program to analyze works (let us limit ourselves to artistic works for the scope of discussion) and produce results with differences from the original data set? Google, back in 2016, had actually held an exhibit for the paintings created by AI software (through advanced Artificial Neural Network) called DeepDream. Artificial Neural Network (ANN) is designed to replicate the human brain’s neural network and is capable of adapting and learning. It takes the whole coding business a step further because the final output is not one coded in but one that the ANN learnt. What DeepDream essentially does is employs ‘Deepstyle’, a technique by means of which knowledge is used to interpret a painting style and transfer it to the uploaded image. Say for example you upload a picture of the sky with clouds and Deepdream sees the human eye or a pizza slice, it would render the image multiple times so that eyes/pizza slice appears in the image. The result is a psychedelic version of the original image input and if information from the exhibit are to be believed, rather expensive ones (with the exhibit raising USD 84,000 in 2 days).
This was not the last time that AI or machine learning was used for art. Some popular examples include — the Portrait of Edmond Belamy (which was sold for USD 432,500), Microsoft’s Rembrandt, a 3D printing technology that uses existing works to create a new Rembrandt or the entire collection of ‘Faceless Portraits Transcending Time’ made by artificial intelligence named AICAN and its creator, Dr. Ahmed Elgammal which in my view seem like a crossover between abstract and classic/renaissance art. Why stop at these instances when there is also a homegrown AI artist? Yes, you read it right. There is an Indian AI as well that makes art — RAGHAV (Robust Artificially intelligent Graphics and Art Visualizer) developed by one Raghav Gupta and owned by Mr. Ankit Sahni (and we’ll discuss it more in a bit).
Sure, now you have taught the AI how to paint (and express itself) but the next question that plagues you is who actually owns the copyright over the work? Let’s back up a little. Copyright is that form of intellectual property right granted to creative works. So basically — paintings, songs, movies, sculptures, books, dialogues, photographs, even performances onstage (drama and live music included). So now, art created by AI would also be protected by copyright (because artistic work). Which brings me back to who actually is the author now? Is it the programmer who taught it how to read, the consumer who entered the image or the AI itself? Or would it just be easier to declare that the work has no author? The question regarding authorship I propose is not novel and has plagued the IP community for a while.
The US law copyright law has a requirement of copyrighted works to be original works of authorship and the work to be created by a human being (Chapter 300 — Sections 306 and 313.2) and this is upheld by their court in cases as well (perhaps I shall cover them soon). However, US is not the only country to do so — Spain and Germany are also countries that require the human element for copyright to be attributable. In fact, in the case of Acohs Pty Ltd v Ucorp Pty Ltd, the Australian Federal Court had held that a work generated with the intervention of a computer was not protectable under copyright because it was not produced by a human. But where there are countries that somewhat disallow AI systems to be authors, rendering them as mere tools, there are countries that allow limited protection to AI-based creations such as UK, India, Hong Kong, New Zealand and Ireland. The easiest way to understand this is to look at India’s Copyright Act that under Section 2(d) (definition of an author) explicitly states that for ‘literary, dramatic, musical or artistic work that are computer-generated, the person who causes the work to be created’ would be the author or a similar section in the UK Copyright Act where under section 9 (authorship of works), it is stated that ‘In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.’ Additionally, the UK Copyright Act (under Section 178: Minor Definitions) specifically provides for “computer-generated” work to be a ‘work generated by computer in circumstances such that there is no human author of work’. The obvious implication would be that the section was so drafted to exclude ‘human authorship’ from copyrightable works, provided that the work does have some ‘modicum of creativity’.
If precedents are to be considered, most courts/registries have upheld the author of the work to be the author/owner of the AI program. A recent example of the same would be the decision of Nanshan District People’s Court, Shenzhen, Guangdong Province, China on Shenzhen Tencent v. Shanghai Yingxun. While the Indian judiciary has not yet faced the issue to adjudicate on it but if one does interpret ‘the person who causes the work to be created’, you know it implies the programmer or the person (as in human person) who had significant control over shaping the art and poor AI was just acting on said instructions.
That said, tides are certainly turning with South Africa recognizing the AI ‘Dabus’ as the inventor of a patent (the first country in the world) and AIVA, a music composing AI being recognized by French and Luxembourg author’s right society as a composer (the first in the world to be officially given the status of ‘composer’). Remember how I mentioned RAGHAV? Well, a painting titled ‘Suryast’ inspired by Van Gogh’s style of painting and a template picture input by Mr. Ankit, was very recently granted copyright protection and the copyright office designated RAGAV as a co-author, in what is a first for not just India but also AI for artworks worldwide. Although industry experts suggest that the same would likely be challenged what happens hence is perhaps what time shall tell — maybe there would be judicial proceedings or legislative developments to come to terms with AIs being the next Van Gogh, Raphael, Da Vinci. But for now, watch out for the cool kids on the block. The AI revolution has come alive!