The ‘AI Elephant’ in the Newsroom

“When I talk of AI, I’m reminded of that famous line: ‘There’s an elephant in the room.’ Today, that elephant is AI,” stated Prof. Sunil Saxena, Founder-Director of the AI Media Academy, during a recent discussion at WAN-IFRA’s AI in Media Forum in Bengaluru. This sentiment underscores the growing urgency surrounding AI’s influence on newsroom practices and the long-term sustainability of media businesses.

The forum served as a critical assessment of AI adoption within the publishing industry, revealing a diverse range of approaches and varying levels of comfort with the technology. From full-scale integration to cautious experimentation, Indian news organizations are navigating a complex landscape.

Diverse Strategies: From Augmentation to Preservation of Trust

At The Printers Mysore, publisher of the English-language Deccan Herald and the Kannada-language Prajavani, AI is primarily utilized for technical tasks. Sowbhagyalakshmi K.T., Director of The Printers Mysore, explained that current applications focus on search engine optimization (SEO), data tagging, and coding, largely within the digital and tech teams. While editorial teams exhibit a mix of “resistance and curiosity,” plans are underway to leverage AI for translation, currently in the testing phase.

Sowbhagyalakshmi emphasized a “human sandwich” model, where AI assists journalists but human oversight remains crucial at the beginning and end of the journalistic process. This approach aims to maximize efficiency while preserving editorial control.

In contrast, Collective Newsroom, the BBC’s Indian-language content provider, adopts a far more restrained strategy. Mukesh Sharma, Co-founder & Deputy CEO, affirmed that the organization avoids using AI for content generation, prioritizing the trust associated with the BBC News brand. However, AI is employed for curation, translations, and basic clip editing, always accompanied by clear disclaimers.

Collective Newsroom has also pioneered an innovative application of AI: protecting the identities of journalists operating in countries with restrictive regimes by transforming their voices. This demonstrates a commitment to journalistic safety in challenging environments.

Reuters, on the other hand, is embracing AI with a more “aggressive” stance. Tresa Sherin Morera, Senior Editor, Publishing & Production at Thomson Reuters, stated their goal is to “augment” existing capabilities. AI has been integrated into their “Leon” content management system (CMS) for proofreading and multimedia asset packaging.

Finally, Manorama Online prioritizes a “human touch.” Santhosh George Jacob, Coordinating Editor, stressed that every stage of production requires human supervision and oversight before publication.

The Multilingual Challenge and the Search Traffic Dilemma

India’s diverse linguistic landscape presents unique challenges for AI implementation. While AI tools for English and Hindi are improving, a performance gap remains for other Indian languages due to limited training data. The preservation of linguistic diversity is crucial, and AI development must reflect this.

Prof. Saxena highlighted that regional language presses are often more attuned to local nuances and complexities – such as caste, religion, and gender – than national or international English-language content used to train many AI models. This local expertise is invaluable.

Interestingly, publishers focusing on local languages appear somewhat shielded from the decline in search traffic experienced by those relying heavily on English content. Manorama Online, with its predominantly Malayalam content, hasn’t been significantly impacted, while Deccan Herald has observed a decrease, likely due to AI overviews and chatbots providing direct answers instead of directing users to websites. This trend underscores the evolving dynamics of search engine optimization in the age of AI. Google’s recent updates regarding AI-powered search results are further influencing this landscape.

Building – or Eroding – Trust in the Age of AI?

The central question posed by Prof. Saxena – “In five years, will AI increase trust in media, decrease it, or make no difference?” – sparked a lively debate. The consensus leaned towards optimism, with panelists emphasizing AI’s potential as an ally in building trust.

Sowbhagyalakshmi K.T. believes AI can enhance credibility by assisting journalists with complex tasks, such as analyzing extensive legal documents. Sharma stressed the importance of newsroom leadership in strategically deploying AI, focusing on tasks like data scraping and content enhancement. He cautioned against viewing AI solely as a cost-cutting measure, advocating for its use in safeguarding journalistic interests and empowering reporters.

Morera emphasized the need for ongoing discussion about how AI can improve journalism responsibly, while Jacob acknowledged AI’s nascent stage but predicted its future power, urging publishers to “tame this beast” while upholding core journalistic values. What role will ethical considerations play as AI becomes more sophisticated? And how can news organizations proactively address potential biases embedded within AI algorithms?

Frequently Asked Questions About AI in News