LSE - Small Logo
LSE - Small Logo

Charlie Beckett

May 20th, 2020

The ultimate data story: driving a global newsroom with AI during the COVID crisis

0 comments | 21 shares

Estimated reading time: 10 minutes

Charlie Beckett

May 20th, 2020

The ultimate data story: driving a global newsroom with AI during the COVID crisis

0 comments | 21 shares

Estimated reading time: 10 minutes

Newsrooms around the world are facing an unprecedented challenge in reporting on coronavirus. It’s a huge, complex, fast-moving story and it comes at a time when news organisations – like other businesses – face logistical and financial problems.

As part of our JournalismAI project, over the last year we’ve been looking at the potential of AI-driven technologies in the newsroom. But what role can they play in coping with these dramatic new circumstances?

In this article Chris Collins, Senior Executive Editor at Bloomberg News based in New York, explains to Polis Director Charlie Beckett how they have included AI technologies in their new ways of working.


 

JournalismAI: What are the key editorial challenges you’re dealing with in trying to cover this story?

Chris: This is the biggest story that many of us have ever had to cover, and the most challenging. The word that we keep hearing from business people is ‘unprecedented’ and it’s the same for us. Some of the challenges are obvious: the volume is most noticeable. It’s a general news story, and a health story, and a financial story – and it’s suddenly coming at you from every direction, which creates a logistical challenge. Setting up home working while dealing with the increased story flow at the same time was not the easiest thing we’ve ever done.

News is about being able to cover not only surprises, but also scheduled news events like earnings releases and government economic releases. Now, suddenly everyone and everything has gone off calendar so the firehose of news is that much more intense. For instance, in one week we covered nearly forty central bank actions. In normal times these are mostly scheduled and we can plan coverage accordingly and ahead of time.

Every corner of the global newsroom has taken a part in telling this story – markets, credit, health care, energy, commodities, finance, politics – and so many more areas.  As with other news organizations, we took a whole newsroom of 2700 journalists and analysts and marshalled it behind one story. It shows the benefits of being nimble, as is an exercise in collaboration.

What are the unique audience and data issues when it comes to coronavirus?

We have a whole world of readers who are suffering through this crisis. How do you find the right way to inform them, to keep the news interesting, but package it in the right way? We have been innovating in this area, with new formats, new story concepts, and different forms of storytelling across our platform. That’s not a new challenge but it’s much more urgent that you change and get it right.

This is also the ultimate data story. Not only do we have infection numbers and health metrics across every country, region and city, but we also have markets data, economic data, corporate data and economic data, which we have to track and understand. From a technical perspective we have built new automated stories and functionality to track the virus numbers in real time: we have virus wraps rounding up companies that have referenced the impact of the coronavirus outbreak on their business, we built a global coronavirus tracker with the most confirmed cases that is updated in multiple languages, and many other examples.

How are you adapting your tools, as well as building new things, in this environment?

Some of the technology we developed in recent years, from news gathering to publication tools, have worked extremely well during this period and helped our journalists stay ahead. For instance, we have a news detection tool used by many of our journalists that uses AI to sift through sources and catch breaking news on certain topics. That is invaluable at a time like this.

We also started to look at the data around the virus: are there automated stories we can produce? For example, we’ve used the technology to track company statements, to identify and show what all the major companies are saying about coronavirus. When it comes to the recent slew of corporate earnings, you change tack because different metrics become priorities. We’ve trained our models to look beyond the usual metrics and find cash flow or debt levels, which are suddenly more in focus for obvious reasons. You also have to look more at what companies are saying, the words they are using, when it comes to the future. As always, it’s a collaborative exercise and a process by which the journalists are training the tools and the AI to pivot and do different things.

How does the product change?

The readership figures show a huge thirst for this information and there’s been record web traffic, as well as record content users on Bloomberg Terminals and across every news platform. We’ve created more daily compilation stories, as well as a daily coronavirus newsletter and podcast. On the terminals we now have a one-stop landing page with the best news coverage that is mostly automated with some human editorial curation.

On Bloomberg.com, we have a dedicated landing page and have put the stories that are most useful to the public in front of the paywall. TV and radio have been doing extensive 24/7 coverage in all regions with guests appearing remotely. We have a new weekly show on our QuickTake platform. Bloomberg Businessweek has had cover stories related to the coronavirus impact running since February, and the staff has been producing the print magazine entirely remotely for several weeks.

What are the long-term effects of this situation on future strategy?

The technology strategy is there to support the journalistic priorities and has worked well during this period. It is focused on how to make us faster and smarter. How to spot news first, and how to spot trends in data that the reporters can respond to.

When you have a story like COVID-19, with the broader newsroom, you look at the technology plans and say: what are the priorities and are they in the right order? As we’ve discussed, you pivot and focus on the coverage and the products that matter most to the audience right now. 

There is a big opportunity, for example, in the story detection realm. If you look at social media, the volume of the information that is circulating and the amount of ideas and general reporting threads that are emerging, there is a lot more we can do with AI tools. Our newsroom faces the same challenge as the audience: the noise level is so extreme. How do you take this noisy world and find the themes that are important? How do you understand and report how people are feeling and acting and what impact does that have on markets, economics, workflow, and so on? Those are big priorities and technology can help.


Interview by Charlie Beckett, who leads the JournalismAI project at Polis, LSE.

If you want to stay informed about JournalismAI insights and activities, you can sign up for our monthly newsletter.

JournalismAI is a project of Polis, supported by the Google News Initiative.

About the author

Charlie Beckett

Posted In: Guest Blog | JournalismAI