Major news organisations across the world have written an open letter directed at regulators and AI-focused tech companies. They called on lawmakers around the world to frame rules to protect copyright in the use of news content for training generative AI models. They also want regulators to facilitate compensation for publishers for the AI-centric use of their published news content by tech giants. The open letter was signed by Getty Images, Agence France-Presse, Associated Press, European Publishers’ Council, Gannett, Authors Guild, European Pressphoto Agency, National Press Photographers Association, News Media Alliance, and National Writers Union.
Generative AI is a type of artificial intelligence that builds AI-driven products for creating texts, images, and other forms of media. At the heart of the copyright controversy are LLMs (large language models), a type of AI trained on published news content available on the internet.
Using the LLMs, generative AI can churn out content for clients. It is anticipated that LLMs may go on to transform the media industry in India and abroad by churning out multiple forms of news content.
India’s media industry is keeping a close eye as news organisations globally are questioning the rising trend of tech companies feeding published news content into their artificial intelligence (AI) models without seeking the publishers’ permission.
A DNPA spokesperson said “Global developments surrounding AI and LLMs, show that there is a need for India’s media companies to factor it in, as they look for ways to democratise the news-sharing and revenue-sharing relationship with Big Tech companies. We expect upcoming Digital India Act and Competition Bill to address these copy right issues for the digital news media in the light of AI tools being used by Big Tech.”
Publishers seek tie-ups
The open letter also pushed for a mechanism to let publishers negotiate deals with AI companies for the use of their news content. Some of the signatories have already inked agreements with AI firms, permitting them to use their published material for training LLMs. Among them is Associated Press, which struck a deal with ChatGPT-maker OpenAI to permit the use of its archive and also use generative AI to produce news content.
In another significant tie-up, OpenAI signed a $5 million agreement last month with the American Journalism Project. It gives domestic news publishers more access to the tech company’s GPT-4-based API.
Google, too, is open to the idea of forging partnerships with news media companies. Its spokesperson Meghann Farnsworth reportedly said the Silicon Valley giant is “exploring ideas to provide AI-enabled tools to help journalists with their work.”
News Corp, NYT cautious
The global media industry, however, is split over collaborating with Big Tech firms that want to scrape published news content from the internet to train LLMs. News Corp CEO Robert Thomson recently told reporters that generative AI could be “fatally undermining journalism and damaging our societies.”
The New York Times, too, chose not to join the publishers trying to negotiate with tech firms over the use of news content for training AI. NYT’s stance is seen as a blow to various publishers’ drive to collectively counter Alphabet and Microsoft over AI use.
While News Corp aired reservations about AI in journalism, its Australia division is reportedly rolling out 3,000 text news stories a week using generative AI. At the Australia offices, a four-member team called ‘Data Local’ is using AI to write news reports in bulk on domestic developments.
Bloomberg, too, has joined the race to utilise generative AI. It recently launched its own LLM, BloombergGPT. The GPT model is trained to analyse niche financial information and create content specific to the industry. Bloomberg reportedly said its LLM was “trained using a 345-billion token public data set.”
History of AI in news media
Some news publishers have been dabbling with AI models since as far back as the nineties. Thomson Reuters has AI integrated into its delivery stack. More recently, it started using an AI tool, Reuters Tracer, to scan X (formerly Twitter) to produce headlines and news summaries. Also, it has been developing an AI tool, Lynx Insight, to help journalists generate story ideas and produce content.
Media giant CNET says some of its published content is created using an in-house AI tool. The BBC, too, has been using AI for creating news snippets and generating stories related to corporate income.
The India scene and Fletcher’s advice
In India’s journalism circles, too, LLMs have become a key talking point. Some educational Institution are already offering short courses on AI that have LLMs such as ChatGPT in the study material. Since the arrival of OpenAI’s ChatGPT, India’s journalism colleges have started introducing the AI tool in classrooms to enable learners to become sharper in the areas of news gathering, copywriting, and editing.
Paul Fletcher, former Australian Communications Minister, recently alerted stakeholders in India’s media industry to acknowledge AI as the new elephant in the room, even as they seek to democratise the publisher-platform relationship. Fletcher, who famously piloted Australia’s Media Bargaining Code in 2021, made the remark at a webinar hosted by the DNPA (Digital News Publishers Association) last month.