Tech Firms and Content Industry Clash Over AI and Copyright at Ministry of Commerce Event

On June 19 and 20, the Ministry of Commerce and Industry’s Department for Promotion of Industry and Internal Trade (DPIIT) opened a vital dialogue among representatives from major tech firms and the content industry, grappling with a pressing question: Who truly owns the data that fuels artificial intelligence (AI)?

At the heart of the discussions was whether tech companies should have the freedom to mine the vast resources of the internet, comprising copyrighted books, articles, music, images, and videos, to train their AI models. 

This debate has significant implications, as the content industry raised serious concerns about their intellectual property being used without consent. Meanwhile, tech companies argued that effective AI training depends on accessing large quantities of data, much of which is protected by copyright.

Startups, in particular, voiced their frustrations on the first day of meetings, urging DPIIT to level the playing field. They emphasized that they are still in the early stages of developing innovative AI models, while larger corporations have already reaped the benefits of AI training, often dodging strict regulatory scrutiny. 

This disparity raises critical questions about fairness in the digital economy. The Digital News Publishers Association (DNPA) made a compelling statement, asserting, “It is imperative to respect the rights of digital news publishers. 

Utilizing our content without consent for AI training and generative applications, like search assistance, constitutes a direct infringement of copyright.” They championed a framework that guarantees fair compensation for content creators, reinforcing their essential role in the evolving digital landscape.

Concerns over being “strip-mined for data” were echoed by creators attending the stakeholder meetings, highlighting the urgent need for protections against unauthorized exploitation of their work.

A pivotal discussion centered on whether India should permit text and data mining (TDM) under an opt-out framework. This innovative technique enables AI systems to analyze and learn from vast datasets, which can include copyrighted material. 

An alternative approach discussed was a statutory licensing framework that would allow AI training companies to utilize copyrighted works, provided they pay a government-established fee and comply with specific regulations.

The DPIIT is actively seeking industry insights on the copyright status of AI-generated works, an issue at the forefront of the ANI vs. OpenAI case pending in the Delhi High Court, where a lawsuit has been filed against OpenAI for allegedly using ANI’s articles in the training of its models.

A thought-provoking question emerged from a representative of an AI startup: “Who rightfully owns the content generated by AI? If a user invests considerable effort into crafting a prompt that yields a unique output, does that make the user the rightful owner, or should the ownership rest with the creators of the AI model?” 

This inquiry underscores the complexity of ownership in an age dominated by AI. These crucial stakeholder meetings are built upon the foundational work of a nine-member DPIIT committee formed in April, comprised of IP lawyers, industry leaders, IT ministry officials, and academicians. 

Meetings have occurred regularly since May 16, with some members expressing concerns about the lack of AI expertise within the group. As this committee gathers diverse perspectives from the tech and content industries, it is poised to release a working paper that could redefine India’s copyright laws to better address the challenges posed by artificial intelligence. 

The outcome of these discussions could very well shape the future landscape of both technology and content creation in India.