India's Government Seeks to Control AI Training Content Costs

The Proposal for a Central Body to Collect Royalties and Distribute Payments to Creators

India's government is considering a new approach to ensure that artificial intelligence (AI) companies pay for the use of copyrighted content when training their models. This idea was introduced in a working paper released by a committee focused on generative AI and copyright, formed by India’s Department of Promotion of Industry and Internal Trade.

The paper highlights that most AI developers have not paid for the use of copyrighted content, which has sparked a global debate about fair use exemptions under copyright law. The department argues that allowing free access to content, known as the "zero price license model," is not suitable because it could reduce incentives for human creativity and lead to less production of original content over time.

The committee also acknowledges that accessing large volumes of high-quality data is essential for AI development. However, they are concerned that negotiating licenses for this content could result in lengthy discussions and high costs, which might slow down innovation, especially for startups and small businesses.

To address these challenges, the paper proposes a hybrid model with three key elements:

1. Establishment of a Central Organization: The paper suggests creating a nonprofit organization called the Copyright Royalties Collective for AI Training (CRCAT). This organization would be managed by associations of rightsholders and would collect royalties from AI companies.

2. Creation of a Works Database: A database would be set up where content creators can register their works. This database would help identify eligible creators who can receive royalties from CRCAT.

3. Simplified Licensing Procedures: By having a single organization handle royalty collection and distribution, the model aims to make it easier for AI developers to access content while ensuring that creators are fairly compensated.

This model draws inspiration from existing systems in other countries, such as performing rights organizations that collect royalties from venues that play music and distribute them to artists. In Australia, there is a similar scheme that charges royalties for reprints of news and other content, distributing payments to registered creators.

India has set ambitious goals to become a global leader in AI. The government has taken a generally favorable stance toward tech giants operating in the local market. However, these companies continue to push for the right to train their models without first paying for content. Despite this, they are also entering into agreements that cover their ongoing operations.

India presents a unique challenge due to its diverse linguistic landscape, with 22 scheduled languages and eight of those spoken by over 50 million people. The country also has a vast and fragmented media and publishing ecosystem. This proposal could be appealing to big tech companies if New Delhi ensures that royalty payments are attractive enough.

While the financial returns for individual creators may vary, some journalists have reported receiving hundreds or even thousands of dollars annually through similar schemes. This indicates that the proposed system could provide meaningful compensation for creators, depending on the scale and structure of the program.