AI will bring disintermediation and efficiencies to media. LSE's Charlie Beckett discusses how that might play out

Two really smart people - Damian Radcliffe of the University of Oregon and Charlie Beckett with the Polis media think tank at LSE - have come together for a conversation about AI. Charlie has been the driving force behind the JournalismAI project, and this work has been way out ahead of the recent surge in interest in AI spurred by the release of a version of ChatGPT last November that suddenly brought the technology to the larger public.

Two things jumped out at me: disintermediation and efficiencies/overabundance, the last two issues I see being linked. As Charlie points out, disintermediation is a little way down the road. This is the fear that AI-driven chat services will provide people with answers and news so that they don’t need to go to publishers’ sites. It will be another way in which technology gets in the way of publishers’ relationships with their audiences and also the commercial benefits that come from those relationships. The second issue is really at the crux of AI and publishing: How will publishers use it? Will they use AI to free up journalists to do more reporting and create more original, more engaging content experiences? Or will they use the technology to generate more redundant content that simply overwhelms audiences?

And Charlie makes a really good point, if publishers are able to drive efficiencies with AI, will they use the funds to reinvest in the business or simply to enrich investors or buy back their own stock. Goodness knows that many news publishers will have a motivation to prop up their stock values. It’s a clear-headed overview that helps set up the timelines and some possible implications of the technology.

And here is how it is being used in the US, in the state of New Jersey. Questions are raised about a small local publisher using AI to generate news reports from local government meetings. I would guess that it summarises the meeting transcripts. But they aren’t checked, only corrected after the fact, which raises a lot of questions.

As I have said, at Pugpig, we’re currently working on a report about retention, and we touch on the issue of zombies, those subscribers, or known audiences who have become disengaged with your content. Dominic Young says that you don't need to fear them, and based on his business, he thinks a pay-as-you-go option is one way to re-engage with these audiences.

For those audience engagement editors out there, this is a really well-thought post about a local newspaper in the US trying Meta’s new Twitter-esque text social network, Threads.

A few of their takeaways that stood out for me:

• It’s not about breaking news updates.

• A strong community feel, usually published during the afternoon. (And that time has always been solid for social media posting. In fact, we used to see a social spike after 4 when people seemed to be wanting to pass some time before quitting time.)

• For this publication, the posts did best when they focused on local issues and the community.

An interesting look at how Norway’s VG is handling corrections. There are some excellent ideas here about how to increase transparency and ultimately trust.

Archives can be the source of a tremendous number of stories, and it has become one of the trends for a certain class of well-to-do publication with deep archives - think the New York Times and The Atlantic. Now the Wall Street Journal is getting in on the action. This is such a good product idea. Even the small newspapers I used to edit in Wisconsin have their own Throw-back Thursdays. So smart and likely to generate some excellent business.

This is so sad and depressing. A woman working for the BBC covering misinformation and disinformation receives a vast majority of online abuse. Trolls are gonna troll. But when is society going to say enough is enough?

The NFT is definitely cooling down.