Climbing the AI Learning Curve

Some links may help generate a commission that supports The Addition.

Almost all of the media industry, one way or the other, is talking about AI. And quite right too. The technology, whichever way it progresses, will have a huge impact on everything from marketing to journalism and beyond. We’re not talking about nonsense like NFTs here.

My view, as I’ve outlined before, is to focus on AI as a tool, not an end in and of itself. My friend Adam Tinworth made another very strong case for all this in his latest newsletter. He rightly pointed out that AI can bring together information and guess at things.

However, it’s not (yet) reliable enough to be a search engine or fact-checker. Newsrooms and individual journalists using it in this manner are largely doubling their work instead of saving time, as they have to go back and confirm details elsewhere after using an AI.

As previously reported at The Addition, AI doesn’t seem to be working well at newspaper publisher Reach. Back in October, one hack there told me that their AI tool “spews out stuff that a five-year-old would write. You put a link into it from another Reach site… and it will reword it into your style. But it then takes another 15 minutes to put it into English.”

AI Development at the BBC

Recently, The Guardian got hold of a memo in which Deborah Turness, chief executive of BBC News, announced the creation of a new AI department. She wrote:

“We must become ruthlessly focused on understanding our audience needs, on delivering the kind of journalism and content they want, in the places they want it, designed and produced in the shape that they enjoy it,

We must do this on platforms that deliver a best-in-class consumer experience, offering personalisation, recommendation and unique user journeys. And we must deploy AI to support, enable and accelerate our innovation and growth.”

This, at least, does seem to approach AI from the perspective of it being a tool, not the output. For instance, Aunty will apparently use AI to recommend articles likely to be of interest to users. That certainly aligns with Adam’s discussion of using AI to make guesses, not offering facts. I’m sure that such a product will give people some good information and direct them to stories that they want. However, it also risks trapping them in an echo chamber, as social media algorithms currently do.

It’s all a learning curve. One that newsrooms of all kinds are having to climb quickly.