Why robots writing breaking news is a great idea

jennisheppardBlog, Media Leave a Comment

Breaking news! The LA Times is using a computer programme to generate breaking news articles.

Although I imagine there will be some traditionalists in the journalistic community who object fundamentally to this process, I think this is a great idea.

Firstly, it offers a simple, innovative solution to the age old problem of staffing – what if news breaks when there’s no one in the newsroom? The BBC has people in the newsroom round the clock – I used to be one of those crazies doing the night shift – but smaller offices of other organisations cannot always provide such cover.

Secondly, setting up an automated breaking news system sidesteps the sometimes tortuous process of using internal, outdated content production systems to manually publish stories which often have the same image, same format and same content.

Thirdly, freeing journalists up from having to regurgitate basic facts, allows them track down greater detail and get reaction or first person stories to make their story stand out. The system gives the organisation a far greater chance of braking the news ahead of the competition, and gives its journalists a head start on their later detailed coverage of the story.

For those still doubting the real world value of this approach, let’s take the LA Times’ own example of an earthquake. Pretty much every story confirming news of a quake elsewhere is based on an alert from the US Geological Survey. The USGS registers earthquakes happening around the globe, of all magnitudes, and anyone can sign up to receive alerts of earthquakes, tailored to region and magnitude as you wish. I follow these alerts and I know a lot of other journalists who do also.

Almost every single breaking news story about an earthquake originates from the basic information provided by the USGS* – location, strength and time of the quake. It is extremely reliable information – in my years covering earthquakes I have only seen the data be rectified twice – one slight change to the magnitude and one larger change to the depth.

Unless you have the misfortune to be where the quake is happening, you’re unlikely to have more information that early on, so why wouldn’t journalists want to automate that system?

In doing so, we’re also ensuring that news of the earthquake reaches the public as quickly as possible. Which is extremely important to the many normal people who don’t follow USGS alerts – and need to know if their loved one is in danger.

A while ago Twitter launched a great ad campaign, showing a guy receiving a tweet, presumably warning him that a tremor is coming his way. The guy has enough time to lift his coffee off the table before the mild tremor strikes, thus saving any spillages and allowing Twitter to claim it is faster than earthquakes.

In my experience as a journalist, those stories breaking news of an earthquake that are not based on a USGS alert or first-hand experience all originate from social media. However, these are unconfirmed until reports are overwhelming or until we have managed to contact the people in question.

By which time the USGS has usually sent out its alert and the robot report would have been published.

Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.