2 weeks Ago By Simone Lobo
Explainer Briefly Slides A student at O.P. Jindal Global University has filed a petition against the institution in the Punjab and Haryana High Court for allegedly incorrectly declaring his exam submission as AI-generated and failing him for plagiarism.
The petitioner, Kaustubh Shakkarwa r is an IP lawyer with the firm KAS &Co currently pursuing a Masters of Law (LLM) in Intellectual Property and Technology Laws at the Jindal Global Law School. Speaking to Medianama, Shakkarwar denied the accusation that he had indulged in plagiarism or used AI. However, he argued that even if he used AI for his submission, he is not liable, as under copyright law the ‘author’ owns the copyright to content they generate on their computer.
He argued that if he had used AI he would have generated the output and thus owned the copyright. Shakkarwar alleged that the University’s ‘ Unfair Means Committee’ used Turnitin’s AI-detection software, accused him of submitting “88% AI-generated” answers, and failed him for plagiarism. He filed a writ petition in the Punjab and Haryana High Court challenging the university’s decision.
Justice Jasgurpeet Singh Puri listed the matter for further hearing on November 14. Medianama has reached out to O.P.
Jindal Global University for its comments. Arguments in the petition The petitioner argued that the University’s rules do not enforce any laws against plagiarism, noting that the university rules known as the “First ordinance of O.P.
Jindal University” need to define plagiarism, get approval from the Government and then notify in the gazette as per section 32 under the Haryana Private Universities Act, 2006. The petition argued that unless the rules defining plagiarism are notified by the Government, “plagiarism” does not count as an offence. He claimed that O.
P. Jindal Global University failed to follow the legal requirement of publishing its first ordinance in the official gazette, which rendered it unenforceable. Arguments regarding copyright Notably, Shakkarwar made the argument that under Indian copyright law, even if he did use AI it does not constitute as plagiarism.
He cited Section 2(d) of The Copyright Act, 1957 , which states that an “author” of a “computer generated” copyrighted material includes anyone that “causes the work to be created.” Shakkarwar argued that even if he used AI he was the “author” of the content and thus the owner of the copyright. The petitioner’s argument for copyright coincides with his argument disputing allegations of plagiarism against him.
He cited the Supreme Court ruling of ‘R.G Anand vs M/S. Delux Films & Ors’ , that held that plagiarism means – a copyright infringement of a content.
The Supreme Court ruled that, “If on a perusal of the copyrighted work the defendant’s work appears to be a transparent rephrasing; or a copy of a substantial and material part of the original, the charge of plagiarism must stand proved.” Shakkarwar claimed that since he allegedly owns the copyright, he is the “author” of the content and therefore the University cannot penalise him for plagiarism. He also contended that using AI tools like generative AI should not violate copyright law because the intellectual input comes from the person using the AI, not the machine itself.
For this argument he cited the sweat of brow doctrine of IP law that assigns copyright to an author based on effort. He also argued that Generative AI has wide application such as grammatical corrections and paraphrasing, which cannot be regarded as plagiarism. Medianama’s take: AI and false positives and negatives The popularity of Generative AI is likely to bring with it a host of disputes such as these.
As an increasing number of students and professionals take advantage of the easy availability of AI models like OpenAI’s ChatGPT and Claude, scepticism may grow among authority figures. However, are AI plagiarism detection models the best way to tackle this seemingly growing issue? Medianama conducted an experiment, wherein it asked ChatGPT to rewrite a Medianama article and then ran it through four plagiarism detectors. Attached below are the results.
As indicated, results across these models ranged from “appearing human” to “0 % human written.” Plagiarism models such as these may often result in false positives or negatives, which brings its reliability into question especially in cases with serious action. Why the debate on AI and copyright matters This case could have wider implications on the debate on AI and copyright and likely set a national as well as international precedent.
In February, former Minister of State Som Prakash said in Parliament that the existing IP laws of India are “ well-equipped ” to deal with AI-related copyright issues. However, cases like these highlight the gaps in the law that are yet to addressed. Generative AI raises questions on authorship and copyright as AI models are usually trained on large sets of data that are often not attributed.
Opinions are likely to differ across the spectrum. In an opinion piece for Medianama, Sneha Jain a Partner (Litigation) at Saikrishna and Associates, specialising in TMT and New Media issues and Akshat Agrawal is Associate at Saikrishna and Associates, LLM from UC Berkeley and a Copyright expert said, “one needs to be mindful of the fact that even if the act of creating substitutes of human creativity, based on datasets that are exemplars of human creativity, seem “harmful” from the point of view of the copyright owner, more often than not, they are not copyright’s concern, unless the expression is actually copied.” They added that, “Independent creation that is not copied is infact fostered in copyright law as against constrained even when it uses the meta-embedded information within previously produced expressions.
” Similarly, at FICCI Frames 2024 Pravin Anand, Managing Partner at Anand and Anand said that the author of a computer-generated work is the person who caused the work to be created. He said, “the moment you have an AI program or AI-generated work, the person who caused it to be created is almost responsible. All you need to do is identify who gathered the data, who trained the model on that data, and who provided the prompts that led to the generation of the work.
So, that somebody, the person who has the closest link to the generated work, is obviously the person. And so, that question has been given more violations.” Also Read:.
