Will artificial intelligence save publishing?

The notion that artificial intelligence (AI) could actually help journalists who are already worried about the longevity of their careers, and publishers who are battling against shrinking revenues, may seem like a long shot, but Jon Stubley argues you need not fear the robo-journalists.

Two years ago, in his final column to readers (well worth a read) before stepping down as editor of The Guardian, Alan Rusbridger mused on the changes the publishing industry had seen in his 20-year tenure.

“It was a world of known knowns. Twenty years later, we swim in unknown unknowns. We still tell stories in text and pictures, but the words are as likely to be in the form of live blogs as stories. We have learned to use moving pictures as well as stills. We work in audio, interactives, data, graphics and any combination of the above. We distribute our journalism across multiple channels, platforms and devices, including live discussion and debate.”

The phrase, ‘We swim in unknown unknowns’ is even more apt to describe the challenges facing the publishing industry today. Last week’s lunchtime debate on the future of independent journalism at Mumbrella360 discussed some of the challenges and possible solutions and came on the back of the Senate enquiry into its future just a few weeks ago.

Undoubtedly, these are tough times for publishers who must (re)find a workable business model and quality journalism alike. As Rusbridger also noted in his farewell piece, “Twenty years ago, no one asked a newspaper editor about their business model. Now it’s one of the first questions.”

The notion that artificial intelligence (AI) could help the situation, particularly for journalists already concerned about the future of their careers, seems to be drawing a long bow but let’s look at the possibilities.

AI isn’t entirely new to the publishing industry. Publications including AP are already using it to automate quarterly earnings reports and stories about college sporting teams. Not to rub salt in the wounds but apparently the automated stories contain far fewer errors than those written by humans.

Still, while journalists have plenty of things to worry about, it seems being replaced by machines isn’t one of them. In 2015, researchers at Oxford University and Deloitte determined journalists, newspaper and periodical editors had only an 8% chance of being put out of work by robots. The research listed the professions as 285th on a list of 366 jobs that could be at risk of automation.

Robotic relevance
In pre-internet days, publishing houses had two unique competitive advantages: content and distribution. Post-internet, the latter has been significantly eroded. Most people have access to the internet and can access news and content in a plethora of ways.

Finding the right distribution strategy in this brave new world is tricky. We’ve seen this up close this month with the news that the aforementioned Guardian has pulled out of Facebook Instant Articles and Apple News, with The Australian dumping Apple News, both blaming poor revenues for the decision, not to mention ceding control and, crucially, data to Facebook and Apple.

Other channel publishers have been relying on social media to spread their message which certainly works for brand awareness, but requires first-party data to create a truly personalised product or service.

Could AI be the answer to these challenges? France’s Le Monde is already using AI tools such as Echobox to shape its social media strategy. The technology can also be used to A/B test headlines that resonate with audiences and to refine distribution strategies.

Yes, AI can help optimise content distribution but it isn’t going to be a panacea or a replacement for a robust strategy. And the process is still going to require human input.

Getting ahead of the tech curve
What cannot be denied is that AI is here and it’s going to have an impact on almost every single industry in the years to come. For the publishing business that has already been gazumped once by the advent of the internet, acknowledging this trend and attempting to get ahead of it is a smart approach. That’s certainly what The New York Times is doing with its newsroom-based research and development lab Story[X]. Launched in 2016, the lab is looking at three prongs of the business – the newsroom, technology and product design and advertising – as it considers the possibilities of everything from chatbots to virtual reality and AI.

One example of how Story[X] has put AI to work to help reporters is The New York Times’ interactive photo of the US presidential swearing-in ceremony. Each swearing in, a staff member at the publication has the thankless task of manually identifying faces in the crowd but in 2017, a model using Microsoft’s Computer Vision API was trained using photos of various members of Congress. The software did the first pass on the image, then a human took it from there.

When we talk AI and publishing, much of the conversation centres on journalists and the impact this technology will have on their jobs. But as Marc Lavallee who heads up the Story[X] lab notes, advertising is an important part of the mix. Lavallee said: “We’re not going to do that alone. It does require a different kind of partnership with a bunch of different kinds of companies.”

AI is advancing in this space at a rapid pace. From personalised creative to AI-driven in-image advertising, there’s so much opportunity for publishers that follow this thread. And the future looks bright in this regard with a recent study finding 80% of 25 to 34-year-olds are all for AI-powered ads that deliver personalised offers.

Instead of fearing robo-journalists, there’s an opportunity for AI to help the publishing industry and I for one can’t wait to see the myriad ways this can be done that haven’t yet been invented.

Jon Stubley is the VP ANZ  at GumGum


Get the latest media and marketing industry news (and views) direct to your inbox.

Sign up to the free Mumbrella newsletter now.



Sign up to our free daily update to get the latest in media and marketing.