News Publishers Flag Use Of Copyrighted Materials By AI Giants Without Consent

Amid the ongoing dogfight between publishers and AI giants, news portals and content platforms have reportedly flagged the use of their copyrighted data for training large language models (LLMs) without permission. 

As per Hindustan Times, the representatives from the content industry raised the concerns during a two-day long stakeholder consultation organised by a panel under the aegis of the department for promotion of industry and internal trade (DPIIT).

Chaired by DPIIT additional secretary Himani Pande between June 19 and June 20, the meeting saw online news platforms term the “training of AI models on copyrighted data without their consent” as copyright infringement. On the other hand, tech giants argued that their models require massive amounts of data – much of which is copyrighted.

“DNPA (Digital News Publishers Association) firmly believes that utilising the content of digital news publishers, without consent, for AI training and subsequent generative AI applications, such as search assistance and information purposes, constitutes an infringement of copyright,” a representative of the industry body said. 

For context, DNPA’s members include the digital wings of some of the biggest Indian media houses, including The Hindu, Dainik Jagran, Network18, NDTV, Indian Express, Malayala Manorama, ETV, India Today Group, Times Group, among others.

The DNPA representative pitched for the need of a regime that ensures fair compensation for content producers and recognises their rights in the digital landscape. Another participant reportedly flagged that creators are worried about being “strip-mined for data”.

Meanwhile, budding AI startups urged the panel to ensure a level-playing field. Citing their rationale, a representative said that larger companies have already trained their AI models without facing the same level of regulatory scrutiny or restrictions while smaller players are still in the early stages of building their LLMs. 

“Who should be considered the actual owner of content generated by AI? If a user puts significant effort into crafting a prompt that leads to a specific output, does that make the user the rightful owner or does the ownership lie with the creators of the AI model?” a representative for an AI startup reportedly questioned. 

As per the report, the participants also debated if the regulators should permit text and data mining (TDM), with an option to opt out of the framework. For the uninitiated, TDM involves scraping large amounts of data including text and images from the internet to train AI systems. 

Amid all this, the participants pitched the idea of a statutory licensing mechanism for AI training purposes, which could potentially involve AI companies paying a “government-set fees” and following certain rules.

The developments come two months after the Centre reportedly set up a nine-member panel to review the existing copyright law amid a rise in AI-related disputes. The committee will examine legal and policy issues related to AI and their implications on India’s copyright law. 

Following the stakeholder meetings, the panel will soon publish a working paper outlining whether India’s copyright law needs to be updated to address challenges posed by AI.

The Fight Over Copyright

The issue came to fore last year after ANI sued AI juggernaut and ChatGPT maker OpenAI in the Delhi High Court (HC). In its plea, the news agency claimed that the AI chatbot “reproduced verbatim or substantially similar extracts” of its works in response to user prompts. 

The petition also claims that OpenAI exploited ANI’s content for its commercial gain by using the new agency’s content to train its LLMs, adding that the usage of its content by the AI major may cause dilution of its market share. 

In January this year, media outlets, including NDTV, Network18, The Indian Express and Hindustan Times, also joined the lawsuit against OpenAI. Under the banner of DNPA, the news publishers claimed that the AI company’s “conduct” constituted “a clear and present danger to the valuable copyrights” of DNPA members and other outlets.

On top of this, the Federation of Indian Publishers (FIP) also moved the Delhi HC earlier this year, against OpenAI on behalf of all its members, which include Rupa Publications, S Chand and Co, Bloomsbury, Penguin Random House, Cambridge University Press and others.

In February, music labels, including T-Series, Saregarama and Sony, also expressed their willingness to join the ongoing copyright lawsuit against the ChatGPT developer in the Delhi HC.