This year we have seen more publishers expanding their artificial intelligence strategies, so we can only expect this to accelerate in 2019.
As inspiration for your future digital strategy, here are four ways artificial intelligence will be used for news media in 2019.
Hyper-personalisation through machine learning
Machine learning is sometimes just another buzzword in the news media industry, with varying understandings of what it actually means. In our machine learning guide released earlier this year, our AI squad prepared a one sentence definition:
“Machine learning uses algorithms that gradually improve on a task without explicitly being toldJoris Gielen, AI & Software Engineer at Twipe
how,i.e. they ‘learn’ from data.”
In 2019, we expect to see more publishers using machine learning to highly personalise reader experiences. They’ll take their cues from publishers such as Neue Zürcher Zeitung, which has built a flexible paywall personalised to individual readers based on hundreds of criteria. They’ve increased their conversion rate by fivefold, with 2.5% of people who view the personalised payment message converting to paid subscribers.
News UK is also working on a machine learning project right now. “JAMES, your digital butler” uses machine learning to gradually get to know the habits, interests, and preferences of readers. The digital butler exposes readers to relevant content in editions in their preferred formats, channels, times, and frequencies.
This helps to increase reader satisfaction and engagement, while ultimately accelerating subscription growth. JAMES transforms conversion and engagement strategies by moving from segmented to highly individualised interactions with readers.
Eliminating bias and fake news
It goes further than just removing political bias, to also giving impartial headlines. For example, on a recent Facebook story, three headlines were produced:
- Impartial: “Facebook scans things you send on messenger, Mark Zuckerberg admits”
- Negative: “Facebook admits to spying on Messenger, ‘scanning’ private images and links”
- Positive: “Facebook reveals that it scans Messenger for inappropriate content”
Even the images used for negative and positive stories were changed to align with the headlines.
Growing real reader engagement through artificial intelligence
Reader engagement, along with solving the funding crisis, is a key topic of concern in the news media industry. In 2019 we expect to see more publishers using artificial intelligence to grow real reader engagement.
Churn prediction and prevention will be a key use of AI. It will provide key insights into the activity of subscribers and applies a predictive model to access their risk, allowing publishers to take the appropriate actions for each risk group.
“Being able to retain customers is far less costly than acquiring new customers. Readers@Risk will allow you to spot changing behavior in the digital reading habits of your readers so you can prevent them from churning.”Jasmien Lismont, Jr. Project Manager & Data Scientist at Twipe
We also know publishers will use more and more AI on text itself. Hannes Buseyne, AI & Software Engineer at Twipe, wrote a report earlier this year for publishers interested in this, “Can AI Predict Reading Time Based on the Words in the Text?”. The report also includes instructions on how to write a similar algorithm to determine the most (and least) engaging words for your own readers.
“This algorithm gives insight into what kind of subjects readers like. With these insights, publishers can better engage their readers.”Hannes Buseyne, AI & Software Engineer at Twipe
Robot journalism expanding coverage areas
First viewed as a threat to journalists, in 2019 we expect to see a deeper understanding of how robot journalism can help expand news coverage areas. It is clear as well that robots are better journalists than news anchors, as Xinhua, China’s state-run news agency, proved last week with the unveiling of their uncanny AI anchors (see their broadcast here).
The Washington Post’s AI powered bot, Heliograf, helps free up journalists to work on more interesting and complex stories, while also allowing for hyperlocal news coverage, such as local high school sports.
“Jeremy Gilbert, The Washington Post’s director of strategic initiatives
Heliografis creating a new model for hyperlocal coverage. In the past, it would not have been possible for The Post to staff more than a handful of the most significant games each week. Now, we’ll be able to cover any game that we have data for, giving the teams and fans near-instant coverage to read and share.”
In its first year, Heliograf published 850 articles and generated more than 500,000 clicks.
MittMedia, Sweden’s leading local media company, also has a robot journalist: “Homeowners Bot”. This bot writes a short text on every house that is sold in their local markets, identifying an interesting angle (such as the most expensive house sold in the year) and adding an image from Google Streetview.
This is a fully automated process, as the names of buyers are a matter of public record in Sweden. In the first four months, this bot became the most productive “journalist”, producing more than 10,000 articles.
It’s also helped to convert hundreds of users into digital subscribers—as its stories are number one in terms of behind-paywall pageviews.
Media innovation analyst @TwipeMobile
Republished with kind permission of Twipe Mobile