AI successes and cautions in first forum event

Jul 13, 2024 at 12:30 pm by admin


Indian publishers laid out their guidelines and reservations on AI, while noting hopes and achievements during a two-day WAN-Ifra forum in Bengaluru.

“AI should only assist and complement the story and writing,” the Times of India’s Rajeev Batra said, adding words of caution about data security, encryption and access control.

Batra (below), who is chief information officer at Bennett, Coleman & Co, told of the publisher’s home-grown BEANS platform during an opening keynote on the first day. The BEANS initiative provides a range of AI-driven solutions for editorial including generation, editing, rewriting, auto-completion, research, transcription, translation and data analysis.

He warned of the need to avoid training on data that could negatively impact language models. “Securing developmental practices in newsrooms is fundamental,” he said. “Ensure that authorisation checks and balances are in place, and maintain continuous monitoring until your LLM is fully developed.”

His warnings were echoed by Newslaundry product and revenue director Chitranshu Tewari who said newsrooms were right to be sceptical of AI and GenAI, “or any emergent technology, for that matter.

“Eventually, AI is a tool,” he said. “Whether we like it or not, it exists, and it’s up to the industry to optimise it to meet journalistic goals.”

Sannuta Raghu, who leads the AI working group at Scroll Media India, addressed the technology’s strengths and weaknesses: “Its strength is in solving a resource problem for small newsrooms and levelling the playing field for us to try and be in the same zone as a mid-sized publisher,” she said, but warned working with AI was “like training a very enthusiastic intern.

“It doesn’t have any context of institutional knowledge, which will result in mistakes, and that’s the process journalists need to navigate.”

Morning Context managing editor Prince Thomas also warned that “from a fundamental level of journalism training, we have technology that, rather than letting you improve, takes over from you.

“How does a cub journalist then enhance their skills? That's an essential problem right now and will only increase in magnitude.”

Attended by about 100 news media executives, the Indian forum is the first of a series of AI events planned by WAN-Ifra. Speakers included AI experts from publishers including The Quint, The Hindu, India Today, Hindustan Times, The Morning Context, Engadget and Newslaundry.

During the first day, Indian winners from WAN-Ifra’s Digital Media Awards Worldwide competition – The Hindu Group for audience engagement and Newslaundry for most innovative digital product – were also honoured on stage.

Keynoting the second day, E&Y director Parankusha S spoke of critical developments and how AI was shaping the news industry. Dating the start of the AI revolution to 2012, he spoke of “an age of deep learning, enabling us to build AI systems that had multiple functionalities… by leveraging your own private data and building products and solutions on top of it.

“The journalism industry is sitting on a wealth of such data. When we talk of GenAI and its impact on the industry, we need to base our strategies on sustainability,” he said.

The second day included discussions on AI for revenue generation, audience engagement and ethical content. At Amar Ujala, personalisation is propelling subscription goals, yielding a 150 per cent increase in CTR, a 25 per cent increase in reading time, doubling the number of active paid subscribers, and a revenue increase of 35 per cent, products assistant general manager Vikas Shekhawat noted.

Quint co-founder and chief executive Ritu Kapur (pictured) chaired a panel discussion on navigating AI in the newsroom, with members LV Navaneeth (chief executive of the Hindu Group), Deepak Saluja (chief executive of Metropolitan Media Company, a part of the Times Group), Mayura Shreyams Kumar (digital business director of Mathrubhumi Printing and Publishing) and Harsha Mathew (chief resident editor and director of Malayala Manorama).

LV Navaneeth told of the Hindu’s newsroom use of AI to inform them of propensity models, dynamic paywalls and pricing. “We have tried to integrate every content solution into the CMS itself. Similarly, we have also tried to integrate AI solutions to enhance advertising and reader revenue into our subscription management system,” he said.

“One of the critical things is to get your purchase decisions on these large pieces of software right. As long as you can build solutions within those architectures, it is far easier to operate than building several isolated solutions.”

The importance of research in decision-making was also emphasised by Mathrubhumi’s Shreyams Kumar: “When it comes to pace, this is not the first revolutionary change we will be chasing. We must try our best to make AI an inclusive change, not a replacement one.”

MMC’s Saluja said vision and strategy were critical for dealing with AI. “For a vernacular language business such as ours, the models still need to be ready for us to absorb it immediately.” As a group Times was “consistently experimenting and dabbling with AI”, he said.

Malayala Manorama was using AI for classifieds and ‘in memoriam’ ads, Mathew said, but “monetising efforts with AI are still a work in progress” with the focus on improving efficiency and making users more acquainted with the technology.

With thanks to WAN-Ifra India

Pictured above: Nagaraj Nagabushanam, data and analytics vice president at The Hindu

Sections: Digital business