AI and the future of journalism

By Catriona Gilmour,

We seem to be reading a new story every week about the automation of jobs and industries all over the world. No industry seems to be totally future-proof – but how about journalism?

Business Wire UK invited PA editor-in-chief Pete Clifton, technology journalist Holly Brockwell and AP’s Paul Shanley to discuss how automation and technology are affecting the industry today.

How are news agencies and journalists using AI today?

From Holly’s point of view, we tend to forget something is AI once we get used to it. Spam filters, Google news alerts, voice assistants – they are all part of our everyday lives, but few of us think of them as AI applications.

Holly runs the website Gadgette, which focuses on technology news for a female audience. She uses AI tools to track search data, and can tailor her editorial schedule accordingly. No natural language processing yet, although she has had some fun playing around with robot-generated articles based on her previous writing. (Anyone who remembers Botnik’s Harry Potter chapter will get the idea of where this technology is right now).

News agencies’ use of AI has been well publicised over the last few years. PA’s RADAR (Reporters and Data and Robots) project has enabled the production of thousands of stories based on data from the Office of National Statistics (ONS). The AI can mine complex data sets for interesting stories for regional newspapers about crime, births, deaths, marriages, healthcare and more.

AP also deploy data-to-text journalism in their output. Quarterly earnings reports, sports statistics and financial market data can all be the basis of AI driven editorial output. AP has invested in text to video software to allow the production of short videos based on their articles, as well as solutions to improve audio to text tools.

What does this mean for journalists’ jobs?

Any workplace automation begs the question: “what about people’s jobs?”

Thus far, the news agencies’ investment in AI has only served to create jobs by scaling their service to levels that would simply be impossible using manual processes.

Their AI projects are journalist led, as the automation does not have the ability to tell an engaging story.

Our panel were keen to highlight that many of the functions AI can fulfil are the mundane, time-consuming tasks that no one got into journalism to do.

For example, audio-to-text transcription technology can save AP journalists around 18,000 hours a year.

AI that can re-write a press release into house style can allow journalists to spend more time delving deeper into the human side of a story.

Using software to analyse the raw data will create more space for creativity with the story, spending time following additional leads and interviewing contacts. 

We have seen the applications of data driven journalism – so what’s next?

Time will tell which areas of journalism will employ automation in future, but it’s likely that this will include:

  • Automated editing – highlight packages, trailers etc. could be quickly turned around to accompany a text article about a sporting event or industry conference.
  • Photographs – while a story is being produced, AI can search the publications image archive for appropriate multimedia and tag it for SEO purposes.
  • Intersectional news – AI can open up brand new avenues of reporting that we haven’t yet had the time or inclination to research. Taking seemingly unrelated data sets and finding weird and wonderful correlations can create really engaging content. For example, is there any correlation between property prices and the locations of Pokémon Go Pokéstops?
  • Accessibility – subtitles, translations and captions will become better and better, bringing news to an international audience and to audiences with disabilities.
  • Personalisation – AI can use publically available data to personalise and target content at certain audiences. Will this exacerbate the issue of “echo chambers” when it comes to news consumption or do people skip what they don’t want to read anyway?

What’s making our panel apprehensive or even pessimistic?

  • Generic, boring content. All of our speakers were concerned that if content becomes more computer-led, editorial standards could suffer. AI can’t replace creativity or human perspectives.
  • Ethics. Industry leaders have the responsibility to ensure that every story involving AI meets the ethical standards of traditional journalism.
  • Cost cutting. In a climate of financial uncertainty for much of the news media, there are concerns that automation could be applied for short-term cost cutting purposes. The panel were clear that using AI should be to add value, streamline manual processes and increase the scope and scale of journalistic output.

Overall, the panel were optimistic about the future

For news consumers, more areas of public life can be opened up for scrutiny and analysis. Publications will be able to use AI to shine a light on issues that historically there has not been the journalistic resources to cover.

For journalists, technology is removing mundane and repetitive tasks from the role, as well as opening up more avenues to tell interesting stories that would not have been possible before. As readerships become larger and more diverse, AI will allow journalism to scale with the audience.

Thank you again to Pete, Holly and Paul for joining us and providing their insights and expertise. Many thanks as well to our guests for joining us for breakfast. We had some great feedback, and watch this space for our next London event.

Photo by Hitesh Choudhary on Unsplash


Leave a Reply

Your email address will not be published. Required fields are marked *