AI has invaded newsrooms; what’s next?

Have you heard of OkayNWA? 

This newsroom in Northwest Arkansas solely employs AI reporters. 

Yes, you read that right. AI reporters. 

OkayNWA calls itself an AI-powered newsroom where AI-reporters Eva Eventful, Wendy Weather, Cultural Carrie generate short reports on local events. 

Sure, OkayNWA is a one-off. However, the use of AI, especially generative AI in newsrooms, is more widespread than one might expect. And why shouldn’t it be? AI has applications in every function of delivering news to readers/viewers. 

Newsgathering? Yes. News processing? Yes. News distribution? Yes.

And this is why I sat down with a group of news publishing experts at the Diginews Conference 2024 to discuss the varied applications of generative AI in newsrooms. We discussed how some of India’s leading newsrooms are already using it, what policies should be put in place to regulate the use of AI, and much more. 

We found that Indian Express Online, The Economic Times Digital, Medianama, and The Print all use generative AI to summarise stories, create explainers, and generate headlines and sub-headlines, albeit with human oversight. 

Use cases 

Nandagopal Rajan, chief operating officer at Indian Express Online, said that while he does not believe in using AI to generate news from scratch and without an editor clearing the content, his team does use AI to create news capsules known as “Express Shorts” in multiple languages.

Medianama, which covers digital and tech policy news in India, has three applications of GenAI. A four-minute long news report could be read in the form of an explainer or question and answer style, in an atomised format, or as slides in a presentation deck. Nikhil Pahwa, founder and editor of Medianama, said that “there are many ways to play with AI depending on the user-need that you want to serve. Because news is a form of service to our readers”.   

At Economic Times Digital, generative AI has been integrated into the content management system (CMS) itself. Deepak Ajwani, editor, The Economic Times Digital, said that GenAI should be included in a journalist’s toolkit alongside calculators and Microsoft Excel. “Keep testing to check the level of perfection and use it because it will definitely save your time.” 

Nisheeth Upadhyay, editor (operations), The Print, said his company has an AI use case for news distribution. The AI model would rate stories on a number of parameters and the highest rated story would be shared on social media in order of ranking.   

In addition to all the text-based applications of generative AI there are numerous solutions available for video content. There are AI tools that can generate videos from text or even snip up long videos in shorter ones because — short is the new black.

Suparna Singh, who led NDTV as its president until 2023 and initiated the news channel’s digital journey, has identified areas where AI can help video news production. Her new AI startup — Frammer — can generate short videos for different platforms in a few clicks.  

Doing AI right

While applications are diverse and powerful, all experts agreed on one thing: human oversight is non-negotiable. In essence, the sub-editor will be much in demand in AI-powered newsrooms. However, “the newsrooms that are scaling up on AI are the ones where sub-editors did not have much currency anyway,” said Nandagopal. 

Introducing AI into newsroom workflows needs consideration too. Pahwa recommended starting with multiple small use cases since “the propensity to make mistakes is less in small tasks”. He suggested breaking down tasks into steps and implementing checks and balances every step of the way. “You will see productivity of the organisation will drastically increase.”  

Increased productivity will have an impact on jobs, causing organisations to become leaner. However, Singh argued in favour of humans keeping their jobs, asking, “Can AI take over our jobs? Maybe. Should it? No.” She said that journalists must control the extent to which AI is used and that it is our responsibility as providers of news to do so.   

A challenge with digital news is that only a few stories that are published each day are read. Every newsroom is trying to make more people discover more stories and that is where AI is being deployed. Nandagopal believes that one of the problems with how AI is being used now is that it is “coming out as a business solution and not an editorial solution”.

AI Policy for newsrooms

Upadhyay put it eloquently when he said, “it is important to not just know when to use AI but also, when not to use it.” He wondered if the time saved by a reporter when using AI may be put towards more creative areas of news production/gathering. 

Not everyone had answers to all questions, but had questions worth pondering upon. And thus we came down to discussing a crucial factor of using AI in the newsroom — do’s and don'ts. Because, like we have established, AI has become an indispensable co-pilot to journalists in almost all newsrooms. 

So, Considering a human was reviewing everything the AI generated, what other guidelines would a newsroom follow? 

Singh suggested disclosures to readers when AI develops any kind of content. Nandagopal said disclosures should be made internally to reporting managers whenever someone in the newsroom uses AI for tasks such as cleaning copy. Indian Express Digital currently does not have a policy on AI but has given its team a list of things to avoid. 

Medianama has a set of do’s and don'ts, as well as a set of banned use cases. Pahwa cited an example of a policy that states that if there is a “significant journalistic impact” that is when a disclosure is made to a reader. How do we assess such an impact? Well, that’s another question for every newsroom to answer for itself.

Disclosures could inadvertently assign shame to the user, although this should not have to be the case. “I have never been ashamed of using Merriam-Webster or Collins Dictionary, so no shame should be associated with using AI,” ET’s Ajwani said. 

It was clear that Indian newsrooms are actively using generative AI and building use cases for more applications across a wide range of departments. It is only a matter of time before AI becomes as ordinary as the ever dependable spell check on a text editor. 

Watch the full video for the detailed discussion.   

We would like to thank our sponsors, MGID (Co-Partner) and Jagran New Media (Gold Partner), for their support.