Introduction
To protect advancements in human creativity and technology is one of the subserving features of IPR. The copyright law in particular is guardian of the creative expression of the human intellect. Through the infusion of amassing technological capabilities accompanied with its ability to automate without instructions and adapt with experiences, the technology is becoming increasingly capable of producing independent creative work. This wide usage of IT (information technology) must be based on a strong legal regulation and licensing mechanism. While indulging ourselves in this legal discourse we must imbibe an in-depth understanding of ownership and authorship in the realm of copyright law, which acts as the core foundation of novelty. With the speedy incoming of chat bots and AI tools like ChatGPT which lay its presence in this digitally onented world. We find ourselves off guard as to what this language processing tool may mean for generating creative works and the ensuing legal implications.
An AI generated song, “Heart On My Sleeve”, which simulated voice of pop stars Drake and The Weeknd was posted by a creator known as @ghostwriter, who claimed that the music was generated by a software trained on these musician’s voices. The music has been pulled out from Apple Music, Spotify, Deezer and Tidal and is in the process to be taken down from TikTok and Youtube as well[1] . Another song “Cuff It” featuring Rihanna’s voice generated by AI Chatbot ChatGPT alos came to the fore recently[2] . The clip has been viewed over 850,000 times and must generated revenue for the uploader of the music. An AI generated cover of Drakes “Passionfruit” in the voice of famous singer Kanye West has garnered over 500,000 views on YouTube[3]. An additional instance of a song which potentially infringed copyright is “Daddy’s Car”. This song was created by an AI named Flow Machines and imitates the style of The Beatles[4] . The AI software analyses thousands of songs from a source database to understand how various artists compiose their music. These instances if seen from a legal lens, make various grounds for copyright infringement and thus require ascertainment.
RAGHAV-AI Painting App
The first ever artificial intelligence tool which was recognized by the Indian copyright office was RAGHAV Artificial intelligence painting app as the co-author with the ownership rights given to Ankit Sahni. Sahni commissioned the painting ‘Suryast’ (Diary No. 13646/2020-CO/A)[5] . However, subsequently The Indian Copyright Office issued a withdrawal notice to Ankit Sahni citing Section 2(d)(iii) and (iv) of the Copyright Act, which deal with the definition of ‘author’ in relation to artistic work created.
Mr. Sahni also filed for protection in other jurisdictions like Canada and USA. The AI tool was granted copyright protection by the Canadian Intellectual Rights Office (CIPO 1188619)6 but the same protection was not attained in the USA. This grant of copyright protection in one jurisdiction and rejection in other acts as an inflection point in the discourse of Al-generated creation and its legal undertones. This article is restricted to the subject matter of Al-generated musical works.There are major issues pertaining to it:
• What protections does the AI generated material has?
• Whether rights can be asserted against AI generated material
• Who is the copyright owner in cases of AI generated creations?
• Accountability in AI generated creations?
What is AI and how it can be used to create music?
Artificial intelligence with the help of huge robust data sets and functionality of computer science simulates human intelligence to enable problem solving and various other outputs. It is earmarked for its efficiency and real time solutions. It has two major characteristics. Firstly, adaptivity supported by experience gaining and secondly automation which means it independently makes future decisions. The AI revolution can be credited to rapidly developing machine learning software. When applied to music works, algorithms learn from the input provided by the programmers and after drawing learnings from these data sets generate new work using its neural network.
AI uses various techniques and applications to compose new music. The ability of AI to create new pieces of music using algorithms based upon user preference creates custom music quickly and reduces the time required to do the same. Al-backed music generators[7] like AIVA, Amper Music and jukedeck can make stylized compositions which can be used by artists to base their music upon. Through machine learning models and deep neural networks, AI technologies produce new piece of sound or lyrics dependent upon the data assessed. These Al-tools are of great assistance to both new and seasoned musicians in order to help them create music, with high AI accessibility. AI for composition may be employed by musicians in ways that are more akin to a cooperative, futuristic tool than an imitation-focused computer. However, there is a legal conundrum to cater to it.
Copyright protection for AI creations and issues of ownership
The law may either completely prohibit or ascribe authorship to the creator of the machine. The former is followed by the western jurisdiction. For instance, in the USA, the US Copyright office mandates[8] that it will “register an original work of authorship, provided that the work was created by a human being.” In the landmark decision of Infopaq International A/S V. Danske Dagbaldes Forening[9] , the European Court of Justice held that copyrightability depends on the originality of the work and originality must reflect the “author’s own intellectual creation”. As understood by jurists, for any copyright work to exist, human authorship is a prerequisite. Pertinently, with the refusal of grant of copyright to RAGHAV Artificial intelligence painting app, India is also Positioning itself to the stance of USA Copyright Office
An “author” is defined as “in relation to any literary, dramatic, musical or artistic work which is computer generated, the person who causes the work to be created” in Section 2(d)(vi) of the Indian Copyright Act. Indian officials are still unsure on how to put this definition into practice. It raises the following question:
• Does the legislation identify the programmer who “causes the work to be created” as the developer of the program or its user?
Section 14 of the Copyright Act 1957, defines “copyright” as an exclusive right of the owner to do or authorize someone the doing acts in respect to the work. Section 17 lays down the exceptions where the author shall not be the first owner of the copyright. For example, if the work is produced under contractual obligation on the instruction of the employer, then the employer will be the owner of the work. The “author” is defined under Section 2(d) as:
(i) in relation to a literary or dramatic work, the author of the work;
(ii) in relation to a musical work, the composer;
(iii) in relation to an artistic work other than a photograph, the artist;
(iv) in relation to a photograph, the person taking the photograph;
(v) in relation to a cinematograph film or sound recording, the producer; and
(vi) in relation to any literary, dramatic, musical or artistic work which is computer generated, the person
who causes the work to be created;]
The issue whether AJ can assume the role of an author requires us to examine Section 2(d) and some case laws. In the case of Rupendra Kashyap vs Jiwan Publishing House (P) Ltd[10] , the Delhi High court held that the questions papers under the head of CBSE compiled by a natural person and not an artificial person, in this case the CBSE. Following is a precept of the judgement:
“(6) By virtue of Section 17 of the Copyright Act copyright vests in the author of the work. The question paper is a “work” within the meaning of the Copyright Act. The author of the examination paper is a person who has compiled the questions. The persons who do this compiling, is a natural person, a human being, and not an artificial person. The Central Board of Secondary Education is not a natural person. It can per se be the author of the papers, entitled to claim copyright in the examination papers. The Central Board of Secondary Education would be entitled to claim copyright in the examination papers only if it establishes and proves that it has engaged persons, specifically for the purposes of – preparation of a compilation, known as question paper, with a contract containing a term that vis-a-vis the question paper prepared by the person engaged for that purpose, that person shall not have copyright and that the copyright would vest in the Central Board of instead of the author of the question paper.”
In Tech Plus Media Private Ltd vs Jyoti Janda & Ors, 2014[11], the court observed that a juristic person cannot be author of any copyrighted work:
“20. The plaintiff is a juristic person and is incapable of being the author of any work in which copyright may exist However the plaintiff can be the owner of a copyright under an agreement with the author of the said work.”
The Delhi High Court in 2019 rejected a copyright claim for a list compiled by a computer because of lack of human intervention[12] .The necessity of the author of a copyright being a natural person has been emphasized and elaborated by courts in various judgements. In Eastern Book Company & Ors vs D.B. Modak & Anr[13] , the court said,
The exercise of skill and judgment required to produce the work must not be so trivial that it could be characterized a purely mechanical exercise. The work should be independently created by the author as opposed to copied from the other works and that it possesses at least some minimal degree of creativity.”
The input required to qualify as being artistic in nature must have the contribution of labor of an artistic kind. The element of originality is very essential to obtain copyright. Originality is not defined under the Copyright Act 1957, however after analysing series of judgements it can be discerned that a work will be considered original as long as the author came up with the work independently and it is not heavily taken from other works. AJ is unable to independently reason and make decisions for itself and operates within predetermined parameters. According to Section 52 of the Copyright Act 1957, the copyrighted work may be used for “fair use”, that is for criticism or review, and it does not require consent from the owner of the work. The exception of fair use will be applicable only if the work generated is transformative. But it’s important to realize that not everyone is born with skill. A person develops skill through studying the works of others in their field. The fact that a person has used earlier works as a source does not prevent him from claiming copyright for his own creations. Even if AI is given legal personality, it is ineligible to assert copyright under the Act because the creator must essentially be a human being. Regardless of whether AI will be treated as a legal personality or not, works produced by AI autonomously need to be safeguarded because copyright law seeks to reward creators. In the case of DABUS (Device for the Autonomous Bootstrapping of Unified Sentience)[14] the patent application by the creator, Stephen Thaler, listed the artificial intelligence system as the inventor. The AI system developed two distinct inventions without any human interference The federal court of Australia (in Thaler v Commissioner of Patents[15] ) and the South African Patent Office[16] are the only two jurisdictions out of more than a dozen jurisdictions. where the patent application has been accepted and recognized.
It would be crucial to discuss how the period of protection would be determined when AI is taken into account as a joint author. Section 23 lays down the timeline of protection granted to the work.
And since AI is immortal its copyright would enjoy protection on the datasets of musical examples creates a derivative version of the original work may impinge the original artist’s right to commercialize their work. For the purposes of this commercial right a licensing model must be proposed to attribute compensation and to the original artist. Regarding training their AI models and technologies. the developers in mind the compliance related to licensing and royalties for data acquisition. This will 4a ihe apt compensation to the artists. A system could be proposed wherein the original artist may receive a portion of the revenue generated by AI in the usage of his/her original work.
Ai-generated work may closely resemble the existing repository of artistic work. which will produce infringing work. The liability and accountability in case of copyright infringement by can be two fold. One is to make the person prompting the response liable another is to make the company or the programmer of the algorithm liable. There exists a gap in the law which will require additions and amendments to the present law.
Comparative analysis
UK has recognized Al-generated work without human interference capable of being protected. The Copyright, Patents and the Designs act, 1988 under Section 9 gives copyright to “the person by whom the arrangements necessary for the creation of the work are undertaken”. In the case of AI generated works, it’s far hard to factor at the person/entity with whose aid preparations have been made. AI investors, AI programmers, AI coders, AI users are all accountable for the AI to generate works autonomously. It might be a case of collective possession and it’s far hard to factor at a particular person/entity to be accountable for the introduction of the AI. Given the multitude of entities which may be involved in generating AI content, it becomes complicated to factor out the real owner. By granting the copyright to the human at the back of the AI are covered for fifty years from the date on which the work became autonomously generated via way of means of the AJ. However, the time period of safety is constrained to 50 years in contrast to 70 years of safety to the works created both via way of means of humans alone or with the help of the AI. The United States Copyright Act 1976 protects works created by a human or by a human with the use of artificial intelligence (AI) as a tool or medium in the creation process. In the case of AI-assisted works, copyright is assigned to the human owner who created the work. Compendium of best practices published by US Copyright Office in its latest version state that if the requirement of being ‘human author’ is not satisfied by the AI machine, then its generated content will not be considered copyrightable.[17]
The United States does not protect the work created i dependently by the AI without human interference. This allows the work to go into the public domain as soon as it is created without protection. According to the most recent version of the Compendium of Best Practices published by the United States Copyright Office, the creative works generated independently by the artificial intelligence machines are not “copyrightable” if they fail to meet the human author’s requirement of the copyright office. Human authorship is a condition that must be met in order to obtain copyright protection in the United States. Al-generated works do not have a human authorship requirement. The US Courts currently apply a nexus test to determine whether there is a significant human contribution in the case of Al-generated works.
The Australian approach does not consider AI to be a person, however the in the recent ruling in Thaler v The Commissioner of Patents[18] , the Australian Federal Court ruled that the AI system could be given the status of an inventor for granting patent protection. It is a complicated process to determine who can be considered as the true owner of AI. Given the situation in the US and UK, it.is to be seen where India positions itself in this legal dilemma. If established as a joint author, we can assume that AI will also be considered a co-author in the subsequent works generated by the AI independently without human input.
If the Doctrine of Work Created for Hire is applied then AI may be considered the employer and the AI creator or the natural person so involved as the employee. Taking this interpretation, the employer will be considered the owner for the works generated by its Al employee. Herein it is to be noted that employee, AI, in this scenario cannot give its independent consent.
A royalty-free domain
Presently, AI generated content is royalty-free but if granted the requisite status of ‘author’, on whom will the liability of the payable royalty lie? Will it be the AI creator or the natural person involved who used the AI system or the AI itself? For example, Boomy which is an AI based music streaming service app, offers royalties to users for their creations.[19] But this royalty is payable to the creator and not the AI involved. If AJ is able to garner the status of ‘author’ in the future then royalties will have to distributed and decided in case ofjoint authorship between a human and an AI system. And if the right of copyrighted AI generated content is infringed them to whom the compensation or royalty is to be paid? These are some questions which require a more nuanced legal approach.
Impact on the music industry
For years music has been a source of solace and entertainment for many. Music has the ability to bind people and shapes their identity as well, such that music is culturally relevant in the society. Al-generated music holds the potential to revolutionize the music industry and transform the listening patterns of the people. In 2020, the musician artist Grims predicted that “I feel like we’re in the end of art, human art,” “Once there’s actually AGI (Artificial General Intelligence), they’re gonna be so much better at making art than us”[20]. While the fear that AI might upend the present industry dynamics, many say it’s for good. AI may spur creativity since it has the ability to analyze huge chunks of data in real time. Easily accessible to musicians, AI will democratize technology in a cost-effective manner. Tasks from pattern identification to automation of repetitive tasks, Al can do it all. Infiltration of AI usage in the music industry will help in managing the increased demand and will help identify consumer patterns towards a more personalized music. Disrupting technologies are affecting the employability in almost every industry, music industry is no exception. AI being more time effective and cheap gains more popularity as a service provider, but it has its own share of limitations. AI may only produce music that is functional or generic in nature. Increased Al-generated work will bring about inevitable legal battles within the sphere of copyright law. Legal recourse will also have to be worked out to make AI a fair space to use.
Conclusion
Al-generated content needs attention of policy inclusion in the legal system of the country. There also arises a need to decide whether rights can be asserted against AI generated material in case of copyright infringement and what can be its pursuant accountability. There is a lack of consensus amongst various nations on the extent and level of agreed protection to be given to Al-generated works. Each nation needs to amend and incorporate provisions related to Al-generated work in their national laws. India must take a step ahead and give due consideration to the issue with an intent to balance the ownership rights, involved risks of infringement and implications on the music industry. Following are the issues to be considered:
• Criteria of originality
• Authorship and ownership of Al-generated works
• Immortality of AI under section 23 of The Copyright Act 1957
• Infringement and issue of asserting rights against Al-generated works
Mansi Kapoor, Intern at S.S. Rana & Co. had assisted in the research of this article.
[1] https://www.bbc.com/news/entertainment-arts-65309313
[2] https://www.insider.com/rihanna-ai-cuff-it-cover-legal-nightmare-music-industry-2023-4
[3] https://www.insider.com/rihanna-ai-cuff-it-cover-legal-nightmare-music-industry-2023-4
[4] https://www.flow-machines.com/history/press/daddys-car-song-composed-artificial-intelligence-created-sound-like-beatles/
[5] https://copyright.gov.in/SearchRoc.aspx
[6] https://www.ic.gc.ca/app/opic-cipo/cpyrghts/dtls.do?fileNum=1 188619&type=1 &lang=eng
[7] https://dittomusic.com/en/blog/ai-for-music-production-tools-for-musicians/
[8] https://copyright.gov/comp3/chap300/ch300-copyrightable-authorship.pdf
[9] https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A62008CJ0005
[10] 1994 IAD Delhi 1, 1994 (1) ARBLR 156 Delhi, 53 (1994) DLT 166, 1994 (28) DRJ 286
[11] 2014 (60) PTC 121 (Del)
[12] Navigators Logistics Ltd. vs Kashif Qureshi & Ors on 17September,2018
[13] 101 (2002) DLT 205
[14] https://patents.google.com/patent/US20150379394AI/en
[15] (2021] FCA 879
https://artificialinventor.com/wp-content/uploads/2021/08/Thaler- v-Commissioner-of-Patents-2021-FCA-879.pdf
[16] https://www.managingip.com/article/2a5czh9 | g6c8zwxjcpla8/dabus-south-africa-issues-first-ever-patent-with-ai-inventor
[17] https://www.copyright.gov/comp3/
[18] [2021] FCA 879
https://artificialinventor.com/wp-content/uploads/2021/08/Thaler-v-Commissioner-of-Patents-2021-FCA-879pdf
[19] https://boomy.com/about
[20] https://time.com/5774723/ai-music/