Repetitive manual tasks have long been earmarked for automation. But a new AI programme underway at Reuters suggests that those in the creative industries should be looking over their shoulders, too.
Reuter’s latest move towards what it calls the “Cybernetic Newsroom” is Lynx Insight. The AI tool has been designed by the news organisation to – in the words of Reg Chua, Reuters’ executive editor of data and innovation – “analyse and sift through data tirelessly, at speed, and on demand.”
Read more: Reuters’ AR & VR visions revealed
The machine newsroom
Before Reuters’ Cybernetic Newsroom has journalists the world over considering a career change, it’s important to recognise that, while a significant step, Lynx Insight is by no means a replacement for authoring original content.
“Humans are good are asking the right questions,” writes Chua, “bringing news judgement to bear, and understanding context. Or to put it the opposite way – and very generally – machines write bad stories and journalists struggle with mounds of data.”
The main aim of Lynx Insight is to analyse data with more speed and efficiency than a human can. Based on that analysis, Reuters says the AI system can then suggest stories and even write key sentences.
Reuters suggests that the tool will be “marrying the best of machine capability and human judgement to drive better journalism, rather than asking one to be a second-rate version of the other.”
A slippery slope?
Given that Lynx Insight can analyse data, pitch stories, and even write aspects of them to an extent, its introduction could be viewed as a slippery slope towards machine-dominated newsrooms.
However, Reuters insists that instead of seeking to replace journalists, it is “placing a bet that the future of automation in the newsroom is less around using machines to write stories, than in using machines to mine data, find insights, and present them to journalists.”
The system will start by offering editors simple insights related to the organisation’s coverage of financial markets. The tool will scour structured financial data for newsworthy items, analysing patterns in the data to spot trends in a fraction of the time that it might take a human to do the same.
It’ll then be up to editors and journalists to apply that data in a way that improves Reuters’ output.
“This leverages the smarts of our newsroom, both in asking the right questions of the machines and in evaluating the answers that come back, to drive even better journalism, and much more quickly,” writes Chua.
“It’s an exciting future that increasingly calls on the undisputed power of robots, but never loses sight of the importance of human judgement and journalism.”
Internet of Business says
“Leveraging smarts” aside (a phrase many human editors would rewrite) Reuters is far from alone in using AI in the production of news. Most of the major news agencies are pursuing automation strategies of one sort or another.
The Financial Times has long applied AI algorithms to push readers towards related content and learn their preferences, as have many other publications.
Meanwhile, News UK – parent company of The Sun and The Times – is automating its subs (sub editors) desk, bringing AI closer and closer to the news itself. The company revealed this in a conversation with Internet of Business editor, Chris Middleton last year.
Elsewhere, some organisations have been trialling the automation of news stories from press releases, while PR agencies have been experimenting with generating automated press releases.
With the increased application of SEO tools to online content also in the mix, a very real risk emerges if automation is pushed too far in journalism and the media: machine-generated stories designed to be parsed by other machines, in order to boost Google rankings. In such a world, human beings become almost irrelevant.
IoTBuild is coming to San Francisco, CA on March 27 & 28, 2018 – Sign up to learn all you need to know about building an IoT ecosystem.