“When I talk of AI, I’m reminded of that famous line: ‘There’s an elephant in the room.’ Today, that elephant is AI.“
This was how Prof. Sunil SaxenaFounder-Director of the AI Media Academy, introduced a discussion with leading Indian media leaders about AI‘s influence on newsroom practices and the long-term sustainability of their businesses.
Speaking at WAN-IFRA’s AI in Media Forum in Bengaluru, Saxena said that the AI elephant “is becoming bigger and bigger with every passing day,” with the related questions becoming increasingly urgent.
The session provided a “reality check” on the current status of publishers’ AI adoption and its real impact on their workflows.
At The Printers Mysorethe publisher of titles such as the English-language Deccan Herald and the Kannada-language Prajavani, AI use has mostly focused on SEO, data tagging, and coding. This has mainly been the domain of their digital and tech teams, said Sowbhagyalakshmi KTDirector of The Printers Mysore.
Among their editorial teams, the use of AI has been less extensive, partly due to “a certain amount of resistance and curiosity at the same time,” she said. However, the company is planning to use AI to translate stories across their publications, although she said this use case is still in the testing phase.
As for using AI in journalism, she emphasised the importance of what she termed the “human sandwich” model: combining AI assistance with a journalist’s input at the beginning and end of the journalistic process.
In contrast, Collective Newsroomwhich is the BBC’s sole Indian-language content provider, maintains a “very limited” approach to using AI, and never uses it related to content generation. This is because the BBC News brand “is all about trust,” said Mukesh Sharmathe company’s Co-founder & Deputy CEO.
However, the company has found AI useful for curation, translations, and simple editing of clips, always using clear disclaimers to indicate AI use.
More recently Collective Newsroom has also come up with a somewhat innovative AI use case.
As the BBC operates in countries with authoritarian regimes, journalists can sometimes be at risk if their identity is revealed. Therefore, the company has experimented with the option of using AI to hide the identity of its journalists by transforming their voices.
In contrast, Reuters has taken a more “aggressive” stance toward the technology, said Tresa Sherin MoreraSenior Editor, Publishing & Production at Thomson Reuters.
She said that their goal with AI is to “augment” what the global news agency already does. AI has also been integrated into their “Leon” CMS for proofreading and packaging multimedia assets for their clients around the world.
Finally, at Manorama Online the goal is to use AI with “a human touch,” said Santhosh George Jacobthe title’s Coordinating Editor.
In their case, this means that “every stage of production has to be supervised by a human,” he said. “And before it’s going live, we’ll have to oversee it.”
Crossing the AI language barrier
As India has a particularly vibrant multilingual news media landscape, questions about the effectiveness of AI systems in local languages are central to the country’s news industry.
Although the quality is “gradually picking up,” Sharma said there is still a noticeable performance gap between the AI tools for English and Hindi (the largest non-English language in the country) and those for other Indian languages. This is mostly due to a lack of training data for smaller languages.
Moreover, Saxena said that India’s local-language regional press is “more in touch with the ground realities” when covering complex and nuanced local issues such as caste, religion and gender, as opposed to the national or international English-language content that most language models have been trained on.
Local languages also seem to offer a protection against the decline in search traffic that some publishers are experiencing due to AI answer engines. For example, Jacob said that Manorama Online hasn’t been impacted by such a decline, as most of its content is in Malayalam.
“But talking about the English side, we have been slightly affected,” he said, though changes in Google’s search algorithm could also be a factor here.
Sowbhagyalakshmi K.T. also said that their English-language title Deccan Herald had seen a decline in search traffic, most likely because of AI overviews and chatbots providing direct answers rather than guiding users to website sources.
Can AI boost trust in media?
“In five years, will AI increase trust in media, decrease it, or make no difference?”
This concluding question from Saxena sparked off an animated conversation about the ongoing trust crisis in the news industry and AI’s potential role in addressing it.
Overall, the panellists struck an optimistic tone.
“I think we should look at [AI] as an ally to help us build trust,” said Sowbhagyalakshmi K.T.
“If we use it to assist journalists in the entire newsroom process, I think it can increase trust and our credibility,” she said. For example, AI tools can help effectively analyse several hundred-page court judgements.
The key is having AI and humans work together, as “the newsroom is very important, it cannot be done without the newsroom,” she said.
Sharma emphasised the leader’s role: “It is for the newsroom leadership to decide how they’re going to use AI,” he said, with possible AI uses including “scraping through large datasets, or enhancing the content to build trust.”
Crucially, this involves viewing AI not as a means of cutting costs, but as a way of safeguarding “the interests of journalists and journalism” and freeing up reporters to focus on their core work.
“That’s where I see trust growing,” he said.
“Over the next few years, [declining trust] is what collectively we should work on,” said Tresa Sherin Morera.
“We should discuss how AI is going to help us in doing journalism better, faster and responsibly,” she said, adding that this will open a “path going upward” for trust in media.
However, Jacob pointed out that as AI is still “in its very nascent stages,” it is likely to become many times more powerful in the future, which puts pressure on publishers to “tame this beast.”
“Then, we will be able to use it for more productive things. And the trust factor will definitely go up,” he said.
“But core journalism values, we’ll have to keep them with humans.”