Here’s a question for the sci-fi geeks out there: what happened on August 29, 1997… at 2.14am?
Well, according to Hollywood, that’s the point at which the machines started to take over the planet. If you’re a fan of the Terminator movies, it’s the moment when the AI supercomputer Skynet became ‘self aware’ and decided it didn’t need pesky humans any more.
Cue lots of explosions, Arnie on a motorbike and a situation for mankind that only got worse and worse. As did the films.
Anyway, cut to the 21st century and it seems that some of these predictions weren’t that far off the mark, at least for the news industry.
The bots have already infested social media, now they’ve set their steely gaze on newsrooms.
At the forefront of this is a project called RADAR – which stands for Reporters and Data and Robots – a joint initiative between the Press Association and data news start-up Urbs Media, with funding from Google.
It aims to produce 1,000 automated news stories a day for the regional press, by scraping information about things like local councils, hospitals, police forces and then moulding them into useable, quality copy with sophisticated software.
This work towards useful AI for the media has been going on for a few years and now seems to be coming to fruition.
I first came across one of the directors of RADAR, Alan Renwick, when he was head of strategy at Local World, the regional newspaper group forged from the old Northcliffe and Iliffe empires.
Local World may have been short-lived before being swallowed by Trinity Mirror, but it was genuinely determined to break the mould when it came to reinventing the news business model, and automation was at the top of the agenda. Alan was the enthusiastic standard-bearer of this brave new world.
I recall one meeting in London where editors were left slack-jawed as the level of automation was outlined to them as part of Local Wold’s plans. It didn’t seem realistic. It seemed pie in the sky. It seemed a little Hollywood.
But seeing how far things have moved on with RADAR makes me now realise just how innovative and prescient those Local World plans were.
So, does it work?
Well, one of the weekly newspapers I used to edit, the Royal Sutton Coldfield Observer, was one of many included in initial trials for RADAR content. Choosing this paper actually set the bar quite high for the project, because data isn’t that readily available for the Sutton Coldfield area.
If you’ve never heard of Sutton, it’s a large and affluent suburb of north Birmingham, which now boasts the biggest town council in the country thanks to a devolution deal.
The problem for RADAR was that the only data available was for north Birmingham, and not nearly local enough for the fiercely independent burgers of Royal Sutton Coldfield. So, we only used a handful of the bot-written stories we were sent.
But here’s the thing – it was the data that was the problem, not the bots. Their copy was pretty good. Sterile, perhaps but I was surprised, impressed and probably a little perturbed by how good it was.
Of course, in the world of PR there have been programmes and gizmos that create ‘automated’ press releases for years. These are usually templated affairs where businesses fill in defined sections, before the programme spits out a formatted press release.
The results lack creativity, rely a little too much on the client’s English skills and can be spotted by a newsdesk at a hundred paces.
But if the bots can find a place in the newsroom, it’s only a matter of time before the tech impacts on PR. Column inches could become even more scarce if Metal Mickey is churning out stories too.
I’m not sure where the march of the news bots fits into my theory of the Great Connection, though.
For those uninitiated: we’re all living through the Great Connection, a tumultuous media era that has taken us in just a few decades from four terrestrial TV channels and press monopolies to smart phones, social media and personalised, curated content.
The natural end-game of the Great Connection is for everyone on the planet to be personally connected to their own multi-platform media feed, the democratisation of media.
At the moment the news bots seem to be serving the old media, helping them fill space at lower cost. It’s only when the masses get control of the bots that they’ll truly join the great global link-up.
Like driverless cars, the prospect of AI worries as many people as it excites. That shrinking violet Elon Musk has spoken of his fears of a Skynet-style catastrophe. Mark Zuckerberg thinks predictions of doomsday scenarios are ‘irresponsible’.
A recent report suggests AI will create more new jobs than it costs.
But, putting aside the shades of sci-fi, I don’t think the march of the news bots is any cause for concern.
The copy they create can’t compete (yet) with stories that are carefully crafted by human hands and minds.
Why? The secret to great story telling is emotion.
As a young reporter I was always told: “Ask your interviewee ‘who, why, what, where, when and how’… but then ask how they felt.”
The emotion was the hook of the story, the thing that would resonate with readers and draw them in.
The AI revolution will no doubt produce valuable and reliable content that will help fill the endless column inches and digital pages produced by the media. It may also help fill the democratic deficit by using data to scrutinise public services.
But until the bots can be taught to scrape that most illogical of databases – the full range of human emotions – they’ll remain the newsroom juniors. I don’t think we need Arnie on his motorbike just yet.