Delhi High Court says training ChatGPT on ANI’s news archive is not copyright infringement
On July 24, 2026, the Delhi High Court refused news agency ANI Media’s request for an interim injunction against OpenAI, in the first substantive Indian judicial ruling on whether AI companies can train large language models on copyrighted news content without a licence. Justice Amit Bansal held, at the prima facie stage, that OpenAI’s use of ANI’s articles to train ChatGPT falls within the fair-dealing exception for research under Section 52(1)(a)(i) of the Copyright Act, 1957.
The Court found that ANI had failed to show that ChatGPT memorised or reproduced its work in a manner that was substantially similar to the original. It also observed that requiring AI companies to obtain licences from every content source would make large language model development economically unviable, since such models depend on access to information already in the public domain.
The ruling does not decide the underlying suit, which continues toward a full trial, including ANI’s separate claim that ChatGPT has fabricated stories and wrongly attributed them to the agency.
Anchor drops 18-year-old “First and Only” toothpaste claims against Hindustan Unilever
On July 20, 2026, oral care brand Anchor agreed before the Calcutta High Court to permanently stop using the claim that its toothpaste was the “first and only” of its kind, closing out a comparative advertising dispute with Hindustan Unilever Limited that had run for eighteen years.
The dispute centred on Anchor’s marketing claims for its toothpaste, which HUL argued disparaged competing products, including its own, by implying that no comparable alternative existed in the market. Comparative advertising claims of this kind sit at the intersection of trademark law and the tort of disparagement, requiring companies to substantiate superlative claims or risk a finding that the advertising misleads consumers and unfairly targets competitors’ goodwill.
The resolution, reached through an undertaking recorded before the Court rather than a full trial verdict, brings a rare formal close to litigation that had outlasted several product reformulations and marketing cycles across nearly two decades.
TVS Motor wins protection against “groundless” copyright threats over its engines
On July 24, 2026, the Delhi High Court restrained the holders of two copyright registrations from issuing further threats of copyright infringement proceedings against TVS Motor Company, in a dispute over claims that TVS’s engines, two-wheelers, and three-wheelers infringed the defendants’ registered “literary works.”
TVS had filed suit seeking, among other relief, a declaration that a cease-and-desist notice sent by the defendants was groundless under Section 60 of the Copyright Act, 1957, a provision that lets a party approach the courts when someone threatens infringement proceedings without actually filing suit. Justice Jyoti Singh passed an ad interim order in TVS’s favour.
The case is notable for the underlying claim itself: technical descriptions or specifications relating to vehicle engineering being asserted as copyrighted “literary works,” a category more typically associated with books, articles, or written documentation than mechanical engine design.
WIPO data shows generative AI patent filings nearly tripled in two years
A World Intellectual Property Organization report published July 14, 2026, and still being widely discussed this week, found that generative AI patent families grew from roughly 14,000 in 2023 to over 37,800 in 2025, with more GenAI patents published in 2024 and 2025 combined than in the entire preceding decade.
China leads the field by a wide margin, with inventors publishing more than 43,000 GenAI patent families across 2024 and 2025, a compound annual growth rate of 64 percent. The United States, Japan, and South Korea follow, with the US posting a faster 92 percent growth rate off a smaller base. Japan’s SoftBank emerged as the single largest GenAI patent applicant globally, ahead of Chinese firms Tencent, Ping An, and Baidu.
The report also found large language models have overtaken generative adversarial networks as the most heavily patented GenAI technology, and that patent filers increasingly include non-tech companies in finance, insurance, and infrastructure, not just software firms.
