Computational Creativity: Art and artificial intelligence

The issues of copyright and owning AI-generated art persist with the increased popularity of newly released software such as ‘DALL-E 2’, ‘Midjourney AI’, and ‘Stable Diffusion’.

There are numerous Indian artists such as Mira Felicia Malhotra and Gaurav Ogale who use augmented reality (AR), where audio-visual mediums can portray immersive visual narratives. Another example is Varun Desai, who creates digital installations using code-generated video art and hand-drawn animation. Indian multimedia artist Raghav KK’s work was recently sold for $94,500 at Sotheby’s Burning Man art auction as a non-fungible token (NFT).

However, the purchase of AI-generated art poses unique legal challenges, especially regarding copyright and ownership. Currently, AI-generated art is created autonomously without human creative input.

India’s copyright law attributes authorship to the ‘person who causes the work to be created,’ which becomes debatable in the case of AI-generated artwork. There are no judicial precedents in India that establish whether AI tools can be considered ‘authors’ and the Indian Copyright Office does not provide clear policy guidance on this matter.

In 2020, the Indian Copyright Office granted copyright to works that were co-authored by AI and humans, but later withdrew it, citing an error. The case is still under consideration, and the inclusion of this case as a precedent remains ambiguous since both the author and the AI tools are registered as ‘co-authors.’

This situation raises concerns, given the potential reliance on this precedent by future applicants.

Under US law, intellectual property can be copyrighted only if it results from human creativity. The US Copyright Office presently recognises works authored by humans. Machines and generative AI algorithms are not considered authors, and their output is not eligible for copyright protection.

The US Copyright Office erred in granting copyright in the case of a comic by Kristina Kashtanova. The dispute arose when the artist used ‘Midjourney’ (an AI programme) to generate artwork for the comic, adding to the ambiguity when it comes to AI-generated creations.

The US Copyright Office held that the content of the comic would be protected under copyright, but the images generated by the AI programme would be under public domain. The US Copyright Office also conducted a session recently covering guidance of works that incorporate AI-generated content and set out disclosure requirements for AI-generated components in copyrightable works, with human authorship as a prerequisite for copyrightability.

Other noteworthy organisations are starting to take note of AI’s increasing involvement in art. The Recording Academy, which is behind the prestigious Grammy awards, recently outlined new rules stating that while songs that include AI-generated elements may be nominated, there must be proof of meaningful human contribution and that only human creators would be eligible for a nomination/award.

Meanwhile, the Writers Guild of America and the Screen Actors Guild are yet to reach an agreement with studios over the potential use of AI in writing and acting roles going forward.

The UK law’s take on copyright is on the same lines, stating that the person making the necessary arrangements for creating the work would own the copyright. However, the US, Spain and Germany have clarified that copyright can only exist and be owned by human beings who meet the requirements of originality.

Since the concept of ‘joint authorship’ is not applicable to AI-generated works, it leads to further ambiguity in using AI-generated works for commercial purposes.

Possible Middle-ground?

Many artists assert that the user owns the work generated by AI. But amid legal challenges regarding copyright, ownership and fair use, a recent partnership between Shutterstock and OpenAI’s ‘DALL-E’ may be a potential solution.

Shutterstock’s programme, ‘Contributors Reimbursement Fund’ compensates artists whose images have been generated by DALL-E and sold on their platform. The partnership addressed concerns raised by artists whose original artwork was used as training data for DALL-E, leading to AI-generated images being sold and licensed without their consent or any compensation.

Future partnerships enabling the use of AI-generated works for commercial purposes would allow the growth of computational creativity alongside artistic innovation, while protecting the rights of artists. These partnerships would create a symbiotic relationship between AI technology and human artists, expanding the realm of possibilities for both.

Moreover, these partnerships would provide multimedia artists with enhanced opportunities to showcase and monetise their work. With the rise of online art galleries, immersive AR exhibits and a booming market for NFTs, artists can reach global audiences.

In a relatively short span of time, generative AI has significantly transformed various aspects of our lives, work and creative endeavors. As a result, it has started to challenge conventional notions of ownership, fairness and the fundamental essence of creativity.

Despite the absence of specific regulations, there is growing recognition and appreciation for the works of multimedia artists. It remains to be seen how the regulatory landscape around this space evolves.

Tanu Banerjee is Partner and Roshnek Dhalla is Counsel at Khaitan & Co. Views are personal, and do not represent the stance of this publication.