Integrating AI into newsrooms: experiment, report, measure

Host: Chris Sutcliffe
January 11, 2024

AI is not a pie-in-the-sky dream for media companies: it hasn’t been for years. For close to a decade questions about whether AI will replace or empower journalists have been raised at conferences and in newsrooms.

Over the past few years the conversation has changed. It is now no longer a question of ‘if’ AI will be integrated throughout the newsroom, but ‘when’.

In December, Media Voices collaborated with Media Makers Meet on their Mx3 AI conference in London. The speakers – drawn from across the media industry from recruitment to newsroom – confirmed that AI is already suffusing the industry. But where they offered invaluable insight was around ensuring it is used effectively and with consideration for the business goals and employees of publishers.

As the official report from the event notes: “A key theme throughout… was the need for publishers to act immediately and start experimenting with AI, albeit with clear business goals at front of mind.”

A selected highlight from the episode:

Buy-in and business…..

One of the hardest parts of integrating any new tools within a newsroom is getting buy-in at board level. Whether it’s off-the-shelf tech or proprietary, there are monetary and time costs for implementing them into existing tech stacks and workflow.

Lucky Gunasekara is co-founder and CEO at Miso, which develops AI tools. He explained that demonstrating ROI on AI tech is vital for getting buy-in: “I think AI needs to put more money into the newsroom than it takes out. Otherwise, it’s just a nice experiment. It’s actually not sustainable, and it will eventually get cut, or something else will get cut in service of it, which is probably even worse.

“So being able to apply AI in ways that actually drive more reader engagement, drive more revenue, improve and strengthen or even upgrade the underlying business model” is what is required to bed AI into newsrooms.

The practicalities around demonstrating that ROI are still being determined. Publishers stand to gain from the use of their own data in training language models, for instance, which should help ease the pain for getting buy-in. Charlie Beckett is a journalist and researcher at the London School of Economics. He says: “There is a there is a huge upside for publishers from getting this right. And creating a licensing model that works for developers, that potentially adds a new revenue stream. We all know that we need multiple revenue streams to create sustainable business models. This is potentially a new one.”

The prize for integrating AI into newsrooms effectively is great. But as the panellists and presenters at MX3 AI explained, media companies cannot and should not barrel in without a plan. It requires reassurance across the business, the implementation of reporting processes to measure effectiveness, and a team dedicated to keeping an eye on all aspects of AI for the newsroom.


This episode is the second of a two-part series from Mx3 AI. The first, looking at opportunities, regulation and risks around AI, can be found by searching for ‘Media Voices’ on your podcast app of choice. There is also a corresponding report from the event, written by Mx3, available to download here



Media Voices and Media Makers Meet would like to thank FT Strategies, InsurAds, Labrador CMS, Miso, Sub(x) and Zuora for sponsoring the conference.