One day when I was a journo the deadline had passed and we were killing time the subject turned to our favourite intros.
If you have never been a hack, an intro is the first paragraph of a news story, and it tends to be. Back then, it was handy to have few tried and tested intros that you can pull out of the bag. There is no such thing as writer’s block in journalism only a lack of professionalism.
My own stock intro for minor stories – better known as news in brief or nibs – was the classic ‘X fans are jumping for job because Y’. Fill in the blanks and you have your instant story. Bingo.
During this newsroom chat we agreed that if someone works out a computer programme to write nibs we’ll all be out of a job.
You may not have noticed but there’s been a shift in the landscape of local journalism… and it’s only a matter of time before it filters into public sector comms.
The tool we imagined back on that lazy afternoon in the West Bromwich newsroom that has since closed down is AI. AI has been steadily moving into the newsrooms across the UK. Over the fence in my own profession there are lessons for public sector comms.
We’re maybe used to the big players in journalism talking about how they are using artificial intelligence tools but how about Berrow’s Worcester Journal? Or the Hexham Courant?
Newsquest, who have 250 titles across the UK, now have seven AI-assisted reporters who use the tool to create content. Under the scheme, the human reporter pastes in a comment and some facts and asks the robot tool to come up with a news story. It’s down for the reporter to fact check the piece before sending it along the production process for it to be used in print and online.
“We’ve produced thousands of articles this way,” the relevant Newsquest executive said, “and we haven’t had any major errors reach the publishing stage.”
It’s probably canny that they’re trialled it Off Broadway at some relative backwater.
Over at Reach plc, they’ve been trialling a tool which helps the very Reach art of re-writing a piece for different websites. Recent work I’ve carried out shows the amount of content shared on local newspaper Facebook pages can be classed as ‘local’. This does not feel like reinforcing a strength if the comments are anything to go by.
Across the industry, journalism is looking at how guidelines for titles can be drawn-up. This makes bags of sense. Have a framework so there can be no comeback to the reporter from the newsdesk or from any people in the newspaper’s front office. Only, the people who may be complaining may not be the man objecting to having his name printed in the drink drive court report but it may be a copyright holder of a large website.
Over in the public sector, there is the knowledge that the ICO’s office is ready and willing to investigate what you are doing with AI and how you are doing it.
When AI went wrong
The former journo side of my brain was fascinated at how using AI blew-up in the face of those behind the Bournemouth Observer. This online news site emerged with a full rosta of named staff none of who really existed.
We can say that the fallout was started by rival news outlets angry at an interloper stealing their turf. Or we could say that there was genuine concern at how trust would be eroded once people found out.
But that’s a really good red flag for public sector comms to consider. I’ve sat in enough Full Council meetings to know how this may sound to a councillor whose knowledge of technology comes shaped by the Daily Express. You have one crack it and if you get that wrong you are stuffed.
Not creating efficiencies but new products
One of the key benefits of AI in public sector comms is to streamline the process and to have something like ChatGPT as a ‘spare brain’. Journalism is looking at much more than that. David Caswell, who has advised 40 different newsrooms across the globe, thinks that innovation in journalism is back. So, new products for new times.
That’s interesting because the public sector after a decade of austerity, a global pandemic and the chance of more austerity knows it should be experimenting but actually doesn’t have much head space.
If you’re experimenting, should you label it?
What’s interesting is a check on the two featured Newsquest newspapers online doesn’t show a mention content is created with the help of AI. No doubt the company would point at the publicity the jobs generated. But most of this was within the trade press rather than in the pubs and bus stops of Hexham and Worcester. I’m not sop sure their readers would be so happy.
Indeed the trend is for labelling. Meta are moving towards labelling content created by AI, as does TikTok and the EU are pressing big tech companies to label AI-generated images, audio and video.
Not only that, but UK Government has been really clear that any organisation using AI should be clear how the tools were used. The UK Government’s ‘Generative AI Framework for HM Government’ says that it should be clear that content has been created with the help of AI. While this stops short of badging everything with an AI label this would point as a minimum for declaring policies and frameworks:
“Transparency is a cornerstone of the ethical development, deployment and use of AI systems. A lack of transparency can lead to harmful outcomes, public distrust, a lack of accountability and ability to appeal.
“Public transparency: where possible from a sensitivity and security perspective, you should be open and transparent about your department’s use of generative AI systems to the general public.”
The authentic stands out
For all I can see the usefulness of AI, early experiments with AI tools show a remarkable insincerity. The video voiced by a generic AI by a Midlands council is remarkably insincere. Make a tool with an accent from your patch and you could be onto something.
Only a few weeks back I blogged at how pics of people were the most effective public sector content on Facebook right now. Why? Because people connect with people. People have Facebook accounts and their own networks, too. That’s where the engagement starts from.
Conclusion
Looking over the fence at journalism, many of the same challenges are being explored that also face public sector comms. The only thing is over there is more urgency. Mass journalism is dying in the UK with fewer reporters and fewer titles. Journalism needs to crack AI as a lifeline for the future.
But public sector comms people would be wrong to sit back in complacency. If we are not shaping change we will be changed at. Find out who needs to be looking at all this and shape the discussion. This feels like a report to Full Council if you are in local government. Or something to be signed off, stamped and agreed if you are in another part of the public sector.
The bad news for the public sector is that without guidelines – just as news companies have – then AI won’t safely get off the launchpad.
Setting up social media in the public sector was at times a struggle. It’s going to look like that was a cake walk.
Better get going.