Editors and chief executive officers are prepared to enlist artificial intelligence (AI) and the power of computing to help them deal with the challenges of the digital age. Indeed almost three quarters (71 per cent) of those we surveyed in our annual digital leaders survey, published today (10 January), said they were already looking at AI as a way of creating or distributing content more efficiently.
Given the traditional agenda setting role of editors and journalists, it is surprising that 59 per cent of publisher respondents are considering letting algorithms select stories for users. Even more, that these algorithms might act independently, learning for themselves what might interest a particular user. But most of these early implementations are still trials and journalists remain clear that ultimate decisions will remain under human control.
How the news industry is already using artificial intelligence (AI)
The core focus for many news organisations, as they move from a model based on reach (and ads) to one based on engagement (and subscription or premium ads), is the need to increase the time spent with a particular news brand. That means providing more personally relevant and timely content which is hard to scale without technology like AI.
An example of how this works comes from China where the most successful news app, Toutiao, has built an audience of 120 million with individual engagement times of 74 minutes per day. Newsfeeds are constantly updated based on what its machines have learnt about reading preferences, time spent on an article, and location. Toutiao claims to have a user figured out within 24 hours.
In the UK, even some of the most traditional journalistic cultures are embracing these ideas. A new recommendation service called James, being developed by The Times and Sunday Times for News UK, will aim to learn about individual preferences and automatically personalise each edition in terms of format, time, and frequency. The algorithms will be programmed by humans but will improve over time by the computer itself working to a set of agreed outcomes.
Other experimental algorithms include a project form Johnston Press called Mooding, which aims to harness a reader’s state of mind, profiling their affinity to different types and tones of content. The Swiss quality newspaper Neue Zürcher Zeitung (NZZ) has been working on AI recommendations for around 18 months and is developing algorithms that are not optimised for clicks but to uphold journalistic standards. These combine ideas of general relevance and personal relevance so individuals are not caught in a filter bubble but are still exposed to more content that they might enjoy. Many of these ideas are being supported by Google as part of its Digital News Initiative (DNI), but organisations like NZZ see the infrastructure they have created as supporting their core business going forward.
Personal Assistants are back
Personal Assistants for journalists disappeared in a round of cost-cutting in the mists of time, but they’re making a comeback in the form of AI bots that can manage diaries, organise meetings, and respond to emails. But some go much further.
Replika is an AI assistant that, with a bit of training, picks up your moods, preferences, and mannerisms until it starts to sound like you and think like you. In the future, maybe it could mimic your posts on Twitter and Facebook – and keep doing it while you are asleep.
AI can also help journalists fact-check political claims in real time – possibly even while conducting a live radio or TV interview. Speech to text robo-fact-checkers will allow journalists to ask questions such as ‘is GDP rising?’ and get instant answers.
Full Fact, the UK’s independent fact-checking organisation, is deploying Natural Language Processing (NLP) with previously fact-checked databases of political claims to test this kind of service to develop two products aimed at helping journalists and fact-checking professionals. The first, Live (see picture prototype below), aims to provide subtitles to fact-check television in real time while the second, Trends, monitors repeated claims and determines who’s making them and where they’re being made, allowing fact-checkers to form a better understanding of why a specific claim is being repeated and to better target correction requests. Both products are due to be released by the end of 2018.
Using NLP to fact-check Prime Minister’s Question Time (concept)
Intelligent automation of workflows
In our publisher survey, 91 per cent of respondents cited production efficiency as a very or quite important priority this year. News organisations know they have to do more with less; they have to find ways of making journalists more productive without leading to burn out. Intelligent automation (IA) is one way to square this circle.
As one example, the Press Association (PA) in the UK has been working with Urbs media to deliver hundreds of semi-automated stories for local newspaper clients (see below). First, a journalist finds a story using one or more publicly available datasets (NHS and population data). The journalist then writes a generic story that is versioned automatically by the computer to create multiple bespoke versions for different local publications. Artificial intelligence is used in a relatively simple way to learn and deal with nuances of language such as swapping out ‘in the Isle of Wight’ for ‘on the Isle of Wight’.
Press Association/Urbs Media process for creating local stories with artificial intelligence
Another potential usage of AI is for reporting of live events. Executive editor of Quartz, Zach Seward, recently gave a speech in China at a conference organised by tech giant Tencent. This was turned into a respectable news story by a combination of AI based speech to text software, automatic transcription, and an automated news-writing programme called Dreamwriter. Around 2,500 pieces of news on finance, technology, and sports are created by Dreamwriter daily.
Taming the machine
Artificial intelligence is set to bring huge disruption to many industries including journalism – and it will be resisted by many. There are real questions over jobs that will be displaced, and over how we use this powerful technology to create value rather than just more clicks. But there are also opportunities to provide better and more relevant services.
Intelligent automation can eliminate repetitive tasks and liberate journalists to find and tell better stories. This year marks the start of the process as we start to grapple with a world when man and machine will need to work closer together than ever before.
Nic Newman is author of the Reuters Institute Trends and Predictions 2018. The research explores likely developments with fake news, tech platforms, business models and more. The full report can be downloaded here.
Notes on survey methodology:
194 people completed a closed survey in December 2017. Participants were selected because they held senior positions (editorial, commercial or product) in traditional or digital born publishing companies and were responsible for aspects of digital strategy. Typical job titles included editor-in-chief, CEO, head of digital, head of editorial development, chief product officer, director of multimedia etc. Half of participants were from organisations with a print background (49 per cent), around a third (32 per cent) represented commercial or public service broadcasters, more than one in ten came from digital born media (12 per cent) and a further 7 per cent from B2B companies or news agencies. 29 countries were represented in the survey including the US, Australia, Kenya, the West Indies, Korea, and Japan but the majority (80 per cent) came from European countries such as the UK, France, Germany, Austria, Italy, Spain and Finland.
Free daily newsletter
- What the Solutions Journalism Network learned talking to ~100 newsrooms about SoJo
- Mobile journalists Sabbir Ahmed and Jamil Khan on creating a mobile-first newsroom
- Reuters Institute Digital News Report: 'Overall, we feel quite optimistic about where the industry is heading'
- Deepfakes, disinformation and detection: How can journalists know what is real?
- How artificial intelligence can help solve journalism’s problems