The robots are coming. According to a 2018 paper published by the OECD, “close to one in two jobs are likely to be significantly affected by automation, based on the tasks they involve,” although “the degree of risk varies” by sector and country.
For some publishers, automation is already a reality, as the technology – and the potential it affords – becomes increasingly mainstream.
In a recent survey of 200 editors, CEOs, and digital leaders, the Reuters Institute for the Study of Journalism found that over three-quarters (78%) of their sample “think it is important to invest more in Artificial Intelligence (AI) to help secure the future of journalism.”
“Most see increased personalisation as a critical pathway to the future (73%),” the report notes, highlighting a major strategic driver for many publishers interested in these tools.
Automation, AI, Machine Learning, and other robo-led processes (which, for the sake of semantics, I’m grouping together here) are rapidly becoming increasingly engrained in the publishing business.
With that in in mind, here are seven uses worth highlighting:
1: Writing articles
Content which tends to follow a formula – such as quarterly earnings reports, and coverage of sports results – has been an early target for automation.
Another leading proponent of this technology, the Associated Press (AP), uses algorithms to produce earnings reports for publicly traded companies.
Harvesting data from press releases, analyst reports and stock performance, these short stories are typically the bread and butter of financial reporting.
“Through automation, we’re providing customers with 12 times the corporate earnings stories as before, including for a lot of very small companies that never received much attention,” Lisa Gibbs, AP’s Director of News Partnerships observed in late 2016.
2: Drafting content and headlines
Forbes continues to expand its contributor network at a time when many others are cutting back. Given this, it’s perhaps no surprise that this is an area that the publication is seeking to enhance.
In January, Digiday’s Max Willens wrote about how the publication’s news CMS would suggest “article topics for contributors based on their previous output, headlines based on the sentiment of their pieces and images too”.
“It’s also testing a tool that writes rough versions of articles that contributors can simply polish up, rather than having to write a full story from scratch,” he added.
Editor Randall Lane observed last year how Forbes’ contributor network had been “our key digital driver since we introduced it in 2010.”
“In bringing in hundreds of expert voices, both paid and unpaid, to the Forbes audience, we were able to create a viable model in a field with too few of them, that greatly increased how much and what we could cover, while also allowing us to expand our full-time newsroom. As a result, we were able to not only survive, but thrive, during a digitally disruptive time.”
As the digital media scholar Alfred Hermida has commented, “AI is fast becoming embedded in the fabric of daily interactions, from navigation apps that optimize driving times to product recommendations based on past purchases; from the algorithms filtering Facebook’s news feed to digital assistants like Siri.”
Netflix is an example of a media company which talks openly about some of their activities in this space. For example, the company has shared how it personalizes the artwork that consumers see for shows, based on their viewing history.
“…let’s imagine how the different preferences for cast members might influence the personalization of the artwork for the movie Pulp Fiction. A member who watches many movies featuring Uma Thurman would likely respond positively to the artwork for Pulp Fiction that contains Uma. Meanwhile, a fan of John Travolta may be more interested in watching Pulp Fiction if the artwork features John.”
“Historically,” the company states on its website, “personalization has been the most well-known area, where machine learning powers our recommendation algorithms.”
But, the company’s usage of this technology goes way beyond that, offering transferable lessons for publishers and content creators in the process.
“We’re also using machine learning to help shape our catalog of movies and TV shows by learning characteristics that make content successful. We use it to optimize the production of original movies and TV shows in Netflix’s rapidly growing studio,” they openly admit.
Moreover, AI doesn’t just help with recommendations and investment decisions. It’s also being used to reduce churn; which is essential given Netflix’s business model.
“Because we’re a subscription company, our North Star is really whether people stay a Netflix subscriber over time,” says Justin Basilico, research and engineering director at Netflix. “…And if we can move that with the recommendation algorithm, we know that we are making people happier and making a better user experience and improving their satisfaction,” he told NVIDIA’s AI Podcast last year.
4: Trend Spotting
Analytics tools offer publishers the opportunity to see what content is trending, as well as how audiences are finding that materials (both from on-site and off-site properties). In many instances, this intelligence is then parlayed into other activities such as social promotion (paid and unpaid) or used to tweak headlines and assets such as specific landing pages.
Just this month, Parse.ly shared how their new alerts service can be harnessed by clients to understand “when a story gets more attention than usual, you’ll get an alert and be able to react in real-time.”
“It’s great to be immediately alerted, wherever I’m at,” said Adrian Ruhi, East Regional Growth Editor at the Miami Herald on Parse.ly’s blog. “Knowing when a story is starting to perform really well allows our teams to jump on them to ensure headlines, photo and video placements, and links to other stories are in place.”
Sourcing material from 4,000 partner sites, “users’ newsfeeds are constantly updated based on what its machines have learnt about reading preferences, from things like taps, time spent on an article and location.”
Based on this, “Toutiao claims to have a user figured out within 24 hours,” the Economist said, adding that “its 120m daily readers spend an average of 74 minutes a day on the app—more than almost any other big social platform in or outside China.”
5: Fact checking
Earlier this month a collaboration between Full Fact, Africa Check, Chequeado and the Open Data Institute secured $2 million over three years as one of the twenty projects supported by Google’s AI Impact Challenge; a $25m (£20m) initiative designed to demonstrate how artificial intelligence can address some of society’s most pressing issues.
Through this work, the teams “will use AI to provide trend monitoring and clustering tools to aid fact checkers’ analysis of news and other information. This will give fact-checkers more time to focus on research, analysis, and writing articles that contextualize the news and help all of us make more informed decisions.”
As concerns about the emergence of “deep fake” videos grow (the technology is still pretty nascent, but that will change, and fast) so humans are turning to robots to help detect them.
“Deepfakes can be used in ways that are highly disturbing,” explains John Villasenor, a Nonresident Senior Fellow at the Brookings Institute. “Candidates in a political campaign can be targeted by manipulated videos in which they appear to say things that could harm their chances for election. Deepfakes are also being used to place people in pornographic videos that they in fact had no part in filming.”
“New technologies open up new avenues for both creativity and manipulation,” Alfred Hermida reminds us.
6: Improving workflows
Outlets such as AP and The Washington Post argue that automating certain types of content, can potentially free up journalists to do other work. This include the ability to do a deep dive into large data sets, themselves often powered by machine learning, help with comment moderation (as seen at the New York Times) or to produce stories more quickly.
At Sky News, for example, AI is being deployed to undertake activities such as facial recognition, and automated subtitling “to unburden journalists.”
“What emerges is a technological drama over the potentials of this emerging news technology concerning issues of the future of journalistic labor, the rigid conformity of news compositional forms, and the normative foundation of journalistic authority.”
Nonetheless, despite these legitimate concerns and tensions, there’s no doubt that these types of tools are here to stay.
7: Creating new consumer experiences
Bots have yet to take off as predicted, but there’s still plenty of experimentation and innovation in this space.
Meanwhile, the New York Times used a Facebook Messenger bot, to provide a kind of modern day “choose-your-own-adventure game,” giving audiences a choice of pre-determined responses to questions and comments posed by the bot.
Nick Confessore, who was one of the reporters covering the campaign for the Times would script these conversations every morning.
“This wasn’t the voice of The New York Times. This wasn’t a quote-unquote “objective” experience. And people really, really reacted — they really liked this experience. We got a quarter of a million people to have some sort of an interaction with Nick Confessore that they thought was personal and sort of unique to them.”
Jump ahead to 2019, and arguably this type of functionality is being explored with smart speaker propositions as well as efforts such as the Quartz Bot Studio.
Automation and algorithms already power large amounts of medialand, even if consumers don’t always realise it.
Programmatic advertising, algorithmically driven front pages, eCommerce, sports and financial reporting, are some of these best known areas, but they’re not the only ones. Given the breadth of potential uses, this is a trend that is not only on the rise, but also one which is impossible to ignore.
Image by Arthur40A, via Flickr