The question of AI usage in community journalism and local news
Plus: Journalism Matters Week shows campaigning in journalism is alive and well
Good afternoon,
It’s back to it after half-term week here, although last week included a flying trip to London for the IPSO (Independent Press Standards Organisation) conference.
Below are some thoughts on AI and the community journalism sector and then the usual digest below of hopefully some useful links which have come across my desk in recent weeks.
Many thanks to Kavita Reddi for excellent panel chairing and it was good to hear from Graham Huband, Robert Barman and Charlie Beckett on their thoughts and uses of AI. And I'm always slightly bemused when I'm sat next to Alison Gow - one of the first editors I ever worked for - who always has the sensible and insightful things to say on pretty much anything.
During the panel I shared how we're at very early stages of using any AI for Blog Preston - we've played with ChatGPT but let's just say, it doesn't know Preston like we do. We are using it quite extensively as part of AI-assisted moderation for large and extensive Facebook groups, and it is really powerful for this (otherwise we'd lose a huge amount of hours each week!).
A few things stuck out to me from the discussion:
🎨 Find your usage and work on it, and make sure it's right for your business. Graham shared how DC Thomson are seeing good results from a recolourisation of archive photos project using AI, it's genuinely adding to audience engagement, content creation and the bottom line of the business.
🤖 I do think we'll see a two-tier approach to AI emerge, for independent publishers like Blog Preston and others then it's unlikely there will be the headspace, resources or capacity to train and develop something bespoke. It may be there is a collective or shared effort in this space (or someone clever like Philip John does something!). Larger publishers will be able to train their own robots to be focused on very specific types of content - as Jody Doherty-Cove and Paul Gallagher are doing with how they are using AI. There's also lots of third-party tools which offer the ability to create custom AI tools and workflows, but these will be cost-prohibitive to the majority of independent publishers.
📱 AI-assisted reporting will become commonplace over the coming years, in the same way smartphone adoption by newsrooms, social media usage by journalists has. But how AI is deployed will be different for different roles, in the same way a football writer uses social media, video and audio, differently to how an at-desk live news/breaking news reporter uses it. As Alison referenced, allowing the space for play and discovery by editors and reporters will be crucial - and then the real creativity will come through as people uncover its real uses in the everyday grind of journalism.
🙋 Kavita referenced an interesting piece of research where audiences still very much want to know a person/human has written a piece of content. Is this an opportunity for particularly community-based titles? As Robert referenced, they'd been cautious across KM Group about utilising AI as they needed to work out how it fitted with a very localised more community news group. A lot of independent news publishers will also be treading carefully. If Blog Preston became pumped full of AI-assisted content tomorrow then it would destroy its essence and what gives it uniqueness in the market. Will we see the 'as told to [name of title]' and bylines and ways of writing that make clear a reporter has left their desk to speak to someone/made a phone call become highly prized (both by platforms and readers).
🤑 There's no doubt there's huge uses for AI in back-office functions and efficiency, everyone referenced this and naturally for established media there's a big focus here. But I've seen lots of creative uses from independent publishers too (using an AI tool to do the drudgery around keeping events and what's on listings up to date for example). Time back for doing the reporting which involves picking up the phone or going to meet someone is always going to be welcomed. And how do you deal with the fact original content is being used by the platforms who build the AI platforms but no funds/money (unless you're big enough to broker a deal) is being passed back down the line to those who put the time/effort/resource into creating the content? For smaller, independent publishers, then this feels an impossible task and ask.
There was plenty more discussed besides the above and it was a good and varied discussion. Thanks to fellow panellists and the questions from the audience too.
I’ve also written some broader thoughts, away from just journalism, on the potential of AI and how to harness it for digital content businesses more broadly. That’s over on our Alma blog.
And now some interesting bits which have come across my feed in recent days.
Journalism Matters Week. As well as the Prime Minister weighing in on a positive note to say how important journalism is, there’s also some great campaigning examples from both regional/local titles and national titles showcased in a public vote. They are on the News Media Association site to vote for.
How TikTok is used for journalism. Sophia Smith Galer. Plenty to unpack about how TikTok users see news and publishers on the platform, appears to be a disconnect between traditional news brands being increasingly entertainment-based on the platform while content creators and start-ups are covering news and issues.
The Lead bets on national and local subscription bundle in Substack move. Journalism.co.uk. I have a vested interest as been involved in this, was good to speak to Jacob Grainger and give some background about
and our local portfolio of , , moving to Substack.And that’s a wrap on this week’s digest, hope you’ve enjoyed reading and do spread the word if you think there’s someone who should be signed up and receiving it. Share the link and pop it in their inbox.
All the best for the rest of the week. Keep going.
Ed
Really interesting piece Ed. Lots to think about. Encouraging that there are options for smaller, independent publishers too - especially as they have the potential to move quicker than the much bigger organisations
The entry barrier with AI for local publishers is really accessible I found — with OpenAI they can train their model based on their own style and writing guidelines (just upload a batch of their content for analysis and training). From what I've seen, that's what the bigger publishers did so far (one was a summarising/document analysis bot, another a rewriting bot). If any indy publishers need a hand with this, I can help (free!) P.S.: I'm happy to be corrected if that's not the case!