Summary

The foreword, by Cherilyn Ireton, Executive Director, World Editors Forum (WEF):

The newsrooms around the world that I come into contact with seem to divide into three camps when approaching the use of Artificial Intelligence.

The Pioneers are those well-funded news organisations, deeply invested in the future and technology, which are using data and machine learning at a high level wherever they can. Organisations such as the Wall Street Journal, AP and Bloomberg are setting the standards.

Experimenters are those who have found value in using algorithms to automate processes and decision-making previously done by humans or to augment content. For example, using machine learning to serve video to every story published online, or deciding what news should be on the front page. Here the Nordic publishers in Europe and News24 in South Africa stand out.

Avoiders – those who don't have the human or financial resources, or data, in clean form, to even begin to think about introducing AI into the news value chain, other than perhaps through third party applications used in the digital distribution process.

This report, the second of the World Editors Forum's 2019 Trends in Newsrooms, captures just some of the experiences of both pioneers and experimenters. From the USA to Sweden and Singapore, we have selected case studies that show how AI can benefit resource-challenged newsrooms, by conducting repetitive tasks, augmenting decision making and helping to deliver a more tailored and engaging offering to news audiences.

An invaluable tool for analysing masses of data, documents

Perhaps the strongest argument for journalists to embrace AI is the benefit it can bring to investigations and verification. The speed at which AI applications can sort through and analyse data, at scale, was an unquantifiable asset for the collaborative teams probing the Panama and Paradise Papers. Amid the new reality of deepfake videos, the technology aiding verification of images and video, and other disinformation, is also invaluable.

Many of my newsroom days were spent on a financial newspaper where well-paid journalists would spend hours churning out formulaic company earnings reports, with just a paragraph or sentence or two of original analysis. In the sports department, loads of students would come in on the weekend to compile sports results. Over the last five years, the Pioneers have proven that tasks like those, or weather stories and news items about recurring events, such as earthquakes, can be written successfully, automatically, with no input from journalists, freeing them to do other things. At this level, there seems to be no downside.

Yet fears linger around job losses, machine-instilled bias and a lack of control.

So how do you deal with these issues if you want to begin the AI journey? Read on. We have valuable input from individuals at the forefront of AI – and some ethical dimensions to consider when introducing AI.

The technology is evolving rapidly, and so too experimentation and adoption. So expect more from WAN-IFRA on this subject in the year to come. This report is our second on this subject this year. It builds on "News Automation – The rewards, risks and realities of 'machine journalism'". We hope together they demystify the road ahead.

Case studies

These case studies are included in the report:

• MittMedia, Sweden
• The Business Times, Singapore
• RADAR, UK
• Aftenposten, Norway
• Dagens Næringsliv, Norway
• Tamedia, Switzerland

Interested in becoming a member of the World Editors Forum? Contact us.