Automating the News: How Algorithms Are Rewriting the Media, by Nicholas Diakopoulos

Ask any older journalist which repetitive jobs they hated and they’ll smile thinly and mention election results and sports league tables. All much-read content but absolute pigs to format. Then shiny new technology put the data into perfectly tabbed columns at the press of a button. Grown hacks wept.

Nicholas Diakopoulos’ timely book moves us forward and spells out with absolute clarity just why algorithms are rewriting the media, and where the wins and the potential losses are.

When it comes to huge wins, think investigative journalism. The Pulitzer prizewinning Panama Papers investigation in 2016 harnessed computer power and freed journalists to do what they’re good at – interpreting the findings. Diakopoulos’ mantra is: “Automate what computers do best, let people do the rest.”

What journalists are good at is spotting an angle, being sceptical and explaining. The Swedish soccer site Klackspark shows how straightforward reports can be left to a computer. An automated content production system churns out 100-word factual summaries of matches, but it takes 14 journalists to add human interest to the stories, such as a girl scoring 10 goals in a game.

Such automated systems can also produce pictures and videos. But watch out for fakes. And beware newsbots, too. They may be invaluable for disseminating online information, and offering journalists the chance to add colour to them, but the little blighters can also potentially wreak havoc by spamming people and flooding hashtags.

It looks like the human touch works best when it comes to news distribution and curation. Staff selections of what to read proved more popular than automised click-throughs to the Chicago Tribune from its Facebook page. This could well lead to a new specialism (or beat for the Americans): algorithmic accountability. Journalists would investigate decisions made with algorithms, particularly issues of public interest, transparency and privacy. There’s certainly a strong ethical dimension to the use of algorithms, one that Diakopoulos could have made more explicit.

The author admits in the acknowledgements that he’s a computer scientist. This may account for the book’s sketchy discussion of news values, and Diakopoulos hand-waves away the idea of reporters having a nose for news and knowing when they see it what makes a story. Most journalists will tell you that this sense comes from on-the-job experience and can’t be replicated by a computer.

The book benefits from a number of interviews, largely with people working in the US industry, but Diakopoulos could have challenged some of them more firmly on what the advances mean as cost-saving measures and the potential for redundancies. Ben Welsh, head of the data team at the Los Angeles Times, claims that it is not about money but rather doing stuff they couldn’t usually do. There will certainly be deskilling, upskilling and reskilling, and Diakopoulos makes the point that interns (properly paid, we hope) can do much of the grunt work. Those of us who’ve been around for a while are familiar with reskilling. We’ve adapted to survive.

A punchy conclusion offers up the challenges, not least the issue of training. While data journalism courses are springing up in the UK, you’d struggle to find much crossover between journalism and computer science. So yes, mindset and training changes will be needed, but writing and newsgathering skills will never grow old.

Sharon Wheeler is senior lecturer in journalism and PR at the University of the West of England, author of Feature Writing for Journalists (2009) and co-editor of The Journalistic Imagination: Literary Journalists from Defoe to Capote and Carter (2007).

Automating the News: How Algorithms Are Rewriting the Media
By Nicholas Diakopoulos
Harvard University Press, 328pp, £21.95
ISBN 9780674976986
Published 28 June 2019