The Delhi High Court has dismissed ANI's copyright lawsuit against OpenAI, ruling that using copyrighted news content to train AI models qualifies as 'fair dealing'. This landmark decision allows private companies to use data for commercial AI training under current laws. The ruling marks a major shift in intellectual property interpretation, though experts suggest lawmakers may still need to intervene to address long-term creator compensation.
In a significant development for the technology and media sectors in India, the Delhi High Court has dismissed an interim injunction sought by Asian News International (ANI) against OpenAI. The case focused on allegations that the AI developer used ANI's copyrighted articles to train its ChatGPT model without authorization. The court ruled that such activity falls under the 'fair dealing' provision of India's Copyright Act of 1957, clearing the way for AI firms to continue using publicly available data for training purposes.
Expanding Fair Dealing to Commercial AI
While the court acknowledged that copying articles for AI training technically qualifies as 'reproduction' under the law, it determined that this usage is protected under the fair dealing exception. Traditionally, this exception has been narrowly interpreted to cover individual research or academic purposes. By extending this to private corporate entities and commercial AI development, the court has adopted a wider interpretation than seen in many other global jurisdictions. The judgment further noted that the outputs generated by ChatGPT did not show 'substantial similarity' to the specific articles owned by ANI.
Market and Financial Implications
This ruling creates a distinct environment for AI companies operating in India compared to those in regions like the European Union or the United Kingdom, where copyright exceptions for AI are often restricted to non-commercial or academic research. A key observation from the court was the lack of demonstrable harm to ANI’s market share or subscription revenue. However, the ruling has sparked discussions regarding the future of content licensing. In international markets, particularly the United States, major publishers have successfully secured licensing deals for AI training rights, with some valuations reaching approximately $3,113 per work. By dismissing the claim, the ruling suggests that Indian media companies may currently have limited legal ground to demand automatic licensing fees from AI developers unless new legislation is introduced.
The Path Toward Legislative Reform
Because India's existing copyright laws were written decades before the emergence of generative AI, the court’s decision highlights a gap that many legal experts believe only parliament can fill. While the court has interpreted the existing act to accommodate modern technology, the judgment implicitly underscores the need for specific, updated laws. Countries such as Singapore have already introduced targeted exceptions for computational data analysis, while the European Union has implemented specific rules for text and data mining. Indian lawmakers are currently evaluating how to balance the growth of the domestic AI ecosystem with the need to ensure fair remuneration for content creators. Investors and media entities will likely track future policy updates or new government regulations that could establish a formal framework for licensing and collective revenue sharing in the era of artificial intelligence.