Did those stable lads who ran into the road to gawp at the arrival of the internal combustion engine have any inkling that their world was about to change? After the wonder (it moves on its own!), the laughter (it’s broken down!) and the complacency (it’ll never replace the horse), it took a few years for real understanding to dawn: the livery yard is closing?
A century on, journalists gazing at artificial intelligence are contemplating a similar but faster cycle. It is a wonder to watch a machine take less than a minute to produce 800 words of analysis in correctly spelled proper sentences. And it’s funny that the first submissions to the BJR are in the style of an earnest teenager. As a member of our production team put it: “Reassuring that they read like a GCSE essay!”
But there is less complacency, for that has already been knocked out of most journalists. The advances in technology that revolutionised the media industry have largely destroyed its business model. It’s easy to be a publisher, hard to make money from it. Few feel particularly secure in their jobs. A first generation of machines began writing sports and business reports for newspapers and news agencies a decade ago. They would, we were assured, free journalists to do more creative work. Rather, they have freed some departments of the need for journalists.
We should say at this point that it is characteristic of a solipsist trade that we discuss whether AI will damage our career chances, rather than address whether it will kill off mankind. These are new technologies that get cleverer with every iteration. When tech bosses start to propose regulation, we know there is something serious to worry about.
Unfortunately, most of us lack the technical ability to grasp how profound these developments will be. We prefer to think instead of the remarkable longevity of newspapers in surviving radio, television and the internet, each of which was expected to kill them. When developers tell us they believe AI will introduce a more powerful revolution than that delivered by the internet, we get a glimpse of what is to come.
Journalists tend to have arts rather than science backgrounds, which might explain our taking refuge in anthropomorphism. For now, we feel more comfortable with the latest AI writers when we see them as rivals, newly arrived on the editorial floor. As paranoid journalists have long done in the face of new talent, we address failings rather than strengths. Sure, it works fast and doesn’t take breaks. Yes, it met that deadline and coughed up a further 250 words to fill the extra space. But if it is so bright, why didn’t it ask what we’d pay? And don’t expect it to work in the UK after midday, when all of America is waking up and playing with it.
The truth is that there are many tasks AI can take on in any publishing house or media production company immediately. It will check spellings and sentence structure, sub-edit copy, create scripts, background stories and explanatory features. That’s become basic. Our ‘guest columnist’, ChatGPT, writing pieces for and against the regulation of the press (below), created perfunctory copy.
But the more ideas we fed in, the more it set off to explore and update. So, what is left for journalists? We might reassure ourselves that AI models are currently merely repurposing information already created. They are, if you like, doing a sophisticated “cuts job”. And who but a reporter can record those first pieces of evidence?
Where does the original information come from? We are deluding ourselves if we believe it is not already out there, often in the useful format beloved of social media. These are machines that scour every scrap of available data, and they do it in seconds. But creative thought? Jokes? Sadness? An understanding of the human condition? Don’t be so sure, for this is technology that learns and improves. When the Today programme set AI to write a Shakespearian sonnet earlier this year, it asked a classical actor and an academic to judge the result. How we laughed at the banality of its words. We’re not laughing now.
KF