Nikita Munjal is an IP Innovation Clinic Fellow, a Student Editor with the Intellectual Property Journal, and a third-year JD/MBA Candidate at Osgoode Hall Law School.
Sabrina Macklai is an IPilogue Senior Editor and a 2L JD Candidate at the University of Toronto Faculty of Law.
In September 2021, the IPilogue reported that India’s copyright office recognized the RAGHAV Painting App (“RAGHAV”), an artificial intelligence (“AI”) tool, as an author of the copyright-protected artistic work, Suryast. The work (reproduced above) was “created” using a base dataset of Vincent van Gogh’s Starry Night painting and a photograph taken by Ankit Sahni, the IP lawyer who created RAGHAV and the work’s listed co-author. Like a natural person, AI might use a painting or photograph as inspiration for their work; unlike a natural person, however, AI can “create” work at an exponentially fast pace. AI’s capability to outperform natural persons is just one of the reasons why its authenticity as an author, sole or joint, is controversial. Labelling the work as a “creation”, as opposed to an “output” or “generation”, has been the topic of global debate.
Following registration in India, Mr. Sahni also achieved success in registering RAGHAV as his co-author for Suryast with the Canadian Intellectual Property Office (“CIPO”) in December 2021 (CIPO, registration number 1188619). This registration marks the first time Canada has attributed copyright authorship to a non-human, signaling a victory for stakeholders who firmly support an amendment of Canada’s Copyright Act (“the Act”) to support the changing needs of innovators and consumers in a high-tech world. Recognizing AI as an author may spark further investment, innovation, and creativity in the Canadian AI sector.
On the contrary, some scholars, including Osgoode Hall Professor Carys Craig, expressed disdain over CIPO registering AI as an author before the government released the conclusions of their public consultation on a modern copyright framework for AI and the Internet of Things (IoT). Concerned stakeholders had from July to September 2021 to submit evidence on whether and how amendments to the Act should be made to achieve its underlying policy objectives while ensuring that Canada’s economy “takes advantage of the opportunities ahead.” The consultation paper discussed three possible approaches for recognizing AI authorship in the Act:
- Attribute authorship of AI-generated works to the person who arranged for the work to be created;
- Clarify that copyright and authorship applies only to works generated by humans (i.e., requiring some human participation for AI-generated works to receive authorship); and
- Create a new set and unique set of rights for AI-generated works.
It seems that CIPO’s registration of Suryast signals the Canadian government’s enthusiasm for the second proposed framework, since RAGHAV is a listed co-author along with its human counterpart Mr. Sahni. However, since the submissions have yet to be publicly shared, some find that this registration amounted to CIPO “jumping the gun” and undermines the purpose of running a public consultation.
While recognizing AI as an author can lead to further innovation in AI-generated works, many drawbacks exist. Some of the most notable arguments are included in a joint submission on the public consultation by 14 Canadian IP scholars. They recommended against recognizing AI as an author and argued that AI-generated works should remain in the public domain. There are technical arguments that the language of the Act implies human authorship and that AI-generated works cannot meet the threshold of “originality” required for copyright subsistence. Further, scholars emphasize that “giving copyright to AI-generated outputs serves none of the [public interest] purposes of copyright protection.” As the Supreme Court of Canada noted in Théberge v Galerie d’Art du Petit Champlain Inc, copyright is usually presented as “a balance between promoting the public interest in the encouragement and dissemination of works of the arts and intellectual and obtaining a just reward for the creator” (at paras 11-12). Where a work lacks significant human involvement (i.e., is truly AI-generated), the scholars argue that no author is denied their “just reward”, as an act of authorship is missing. Similarly, they argue that there is no reason to assume that AI-generated works will be under-produced in the absence of copyright protection, and so the incentive copyright is meant to provide is absent.
It is important to remember that once an applicant files their registration with CIPO, the office conducts a formal check of the details submitted in the application. Neither this oversight process nor the certificate of copyright registration amounts to a guarantee of the legitimacy of ownership or that the originality of the work will remain unchallenged. The lack of critical examination throughout the process is significant and may not be the victory for AI that many proclaim it to be. Theoretically, granting registration imparts onto the AI “author” the same rights and remedies that a human author would receive under the Act. An AI could enforce its copyright if a user is infringing. However, a user could challenge an AI’s copyright-protected work on the grounds that it lacks originality and, therefore, lacks copyright altogether. While this is a hypothetical situation, given the amount of controversy this registration has generated, it would be unsurprising if legal action followed.
Although the reach of this registration is limited, it does showcase the growing uncertainty around how AI interacts with copyright laws. Only time will tell where Canada stands on AI authorship as we await the results of the public consultation. Regardless of the position taken, the government must act urgently to address AI and copyright. These questions only become more complex as technology evolves.