Google is funding a project where 1000 news stories a day will be auto-generated. But US media is already ahead in this game.
Bangalore July 20 2017: From Hollywood classics like "The Front Page", to Robert Redford and Dustin Hoffman unraveling Watergate in "All the President's Men" to our own Seema/Sridevi trying to unmask "Mr India", the journalist pounding away on a typewriter ( or latterly at a desktop PC) is the very stereotype of how news stories are written. Perhaps not for long. The Robots are taking over!
Earlier this month, the UK's national news agency, Press Association ( PA), announced that it had received a $ 800,000 ( Rs 5.2 crore) grant from Google's Digital News Initiative fund for its RADAR -- Reporters And Data And Robots -- project. The PA is trying to see if it can generate up to 30,000 stories a month -- for local and regional consumption -- automatically without human effort. Journalists ( human) will tap mostly large government and local administration data bases for the system to auto-generate stories relating to employment, health, crime etc. So in future if readers in a small British town, read in their morning paper about new job opportunities in the city council or a local health warning or stories of local crime culled from Police press notes -- chances are, no human wrote them; they were machine generated by RADAR.
PA is trying out this robot-reporter jointly with Urbs Media, a startup specializing in data-driven news. The two will deploy a small team of around 5 humans to oversee the automated output.
Things are still in the learning stage: PA expects to launch this machine-generate flood of localized news early in 2018.
Radical as it sounds, it is not the first time wire services have experimented with robot news generation.
The US experience: Even two years ago, the Associated Press in the US, began auto-generating sports news, starting with collegiate baseball. This ensured that every town with a college, got baseball news about its own team. AP used software created by Automated Insights, a company specializing in tapping Big Data to produce news content. The technology was also used for the first time that year to report the quarterly earnings of Apple, within seconds of the information becoming available, and ahead of manually reported news agencies. Automated Insights said its Natural Language Generation (NLG) platform could create 2000 news articles per second, if called to do so.
In 2016, The Washington Post newspaper used its own neo-news tool called Heliograf, to auto-generate bursts of result updates at its news blog, as well as tweets, during the Rio Olympics.
If media organizations can save money on hiring humans as reporters -- they will , where ever possible. Right now only the more mundane news -- long lists of football or racing results or news where the official press release needs minimal editing -- is being auto generated. But who knows how long it will take for owners to expand the role, to reduce the number of living breathing journalists?
"I don't think computer whizz-bangery is going to replace bonafide reporting", said the head of the (UK) National Union of Journalists, Tim Dawson, warning against the trend if it ends up churning out 'third rate stories" which merely "look exciting".
The Machines are slowly creeping into the newsroom. How long and in what areas, will Man prevail?