What TV doesn't, print can't; we deliver.
What TV doesn't, print can't; we deliver.
Pivotal global event unites over 3,000 leaders, policymakers, and tech experts from 190+ countries
NEW DELHI: At the 15th ITU WTSA in New Delhi, researchers examined the legal complexities of AI and copyright. The paper, “Copyright in the Age of Artificial Intelligence,” by Hoshiar Singh and Kiran Sharma, highlighted how AI disrupts traditional copyright frameworks. The discussion emphasised the need for updated legal protections for AI-generated work.
The research highlights how AI has transformed from a tool for simple tasks into a creator of artistic and intellectual works such as literature, music, and visual art. These capabilities raise the fundamental question: Who owns the copyright of AI-generated works? Does it belong to the programmer, the AI itself, or the user who initiates the creation? The paper examined this debate through several legal theories, including the “labour theory” of copyright, which traditionally awards protection based on human effort.
The research focused on ownership and authorship, a cornerstone of copyright law, which has historically been reserved for human creators. However, AI blurs this distinction, as AI-generated works, although based on algorithms and data inputs, exhibit creativity once thought unique to humans. The study examined whether AI-generated works can qualify for copyright protection and, if so, who should be recognized as the author. The research acknowledged that current copyright laws in India, governed by the Copyright Act of 1957, only allow human authorship, assigning the copyright to individuals who commission AI-created works. Globally, the debate is equally unsettled, with countries like Japan and the U.S. wrestling with how to attribute copyright to non-human creators.
ALSO READ: Delhi police probe Khalistani link in Rohini blast, seek Telegram’s help
A critical issue explored in the paper is the use of copyrighted material in AI training. For an AI to generate original content, it must be trained on vast datasets, often composed of copyrighted material. The paper raised concerns over whether this constitutes fair use or infringement, especially if the AI’s output closely resembles the works it was trained on. The research pointed out that this ambiguity leaves room for legal disputes, as seen in recent cases involving AI-generated art and music.
It also provided an overview of how various countries approach the issue of copyright for AI-generated works. Countries like China grant copyright protection if humans play a significant role in the creative process, whereas the U.S. limits copyright to works with clear human authorship. The U.K. takes a middle ground, granting authorship to those who set up the AI system. The researchers called for international harmonisation to provide clarity and consistency across borders as AI-generated works continue to proliferate.
ALSO READ- AROHA 2024: First international Ayurveda conference kicks off
The research underscored several unresolved questions, including the lack of a global framework for regulating the copyright of AI-generated content. It also noted that the current legal framework may fail to recognize AI as an independent creator, as AI lacks the human-like attributes traditionally associated with authorship, such as creativity, emotion, and intent. Singh and Sharma warned that these gaps leave creators, programmers, and companies in legal limbo regarding ownership and liability.
The research presented at ITU WTSA 2024 emphasised the need for a new copyright framework that adapts to the realities of AI-driven creativity. The authors advocated for a balanced approach that protects human creators while acknowledging the unique role AI plays in content generation. They also called for clearer guidelines on fair use in AI training, as well as international cooperation to establish a unified legal framework for AI-generated works.
