That's a clickbait headline that isn't appropriate for the article, the original is "The Guardian’s approach to generative AI".
They seem to be only using it in the background - "helping journalists interrogate large data sets, assisting colleagues through corrections or suggestions, creating ideas for marketing campaigns, or reducing the bureaucracy of time-consuming business processes"
If and when they use editorially, the seem to be checking and moderating with human oversight and "will be open with readers" when they do so
I know several professional journalists, I even happen to employ (a former) one.
None of them are using GenAI to write their articles. To help with research, topic idea and generation, maybe getting some options for how to rewrite a passage or headline, yes. All great, used right it helps them be more efficient and improves the end result.
I have a very amateur background in journalism myself (edited my college and high school newspapers, took a class or two). I can smell excessive ChatGPT abuse on an article. It's like how good devs can smell bad code. ChatGPT tends to write articles which regurgitate popular viewpoints from the Internet in an equivocating way that isn't very engaging and contains no unique insights. In other words it's not output that most professional newsrooms would want to publish, it doesn't hook anybody (at least not yet). Current generative models also handle current events poorly. The Guardian are quite good at what they do and abusing e.g. ChatGPT would dilute their brand.
Does that mean they might do something like "hey chatgpt, what percent of people think x has the right policy on y?" and then just publish it as a factual survey?
The actual uneditorialised title is "The Guardian’s approach to generative AI", and the reality is far less sensational than the invented title "The Guardian will now use generative AI" implies.
This honestly seems like a reasonable, sober approach to LLMs that considers how they may serve journalists in a few key ways, whilst retaining them as the ones actually writing the content. Using LLMs to help sift through large datasets and assist the (hopefully sufficiently Church/State separated) marketing team is honestly a great way of staying able to deliver accurate information to readers.
The title (for some reason editorialized from the more fitting "The Guardian’s approach to generative AI") could give a person that doesn't read their release the wrong idea that they may consider utilizing LLMs for article creation in a way that they at this point in time don't seem to intend.
Don't know why the change from the original was made, as this can only lead to confusion and controversy and does not serve anyone (unlike e.g. adding the year something was posted).
Bad HN title for quite a reasonable article saying that they won't use it where there is risk of unreliable facts creeping it but will leverage it for things like interrogating large datasets.
Makes sense to me, generating queries using natural language seems to be one of the less tricky areas to get right.
> We don’t yet know the full impact that these new technologies will have on our society, but we feel sure that trusted media organisations which prioritise intelligent original reporting, uncovering facts, holding the powerful to account, and interrogating ideas will be as important as ever before
I think that this explanation is good, but I also hope that the first phrase will not have a negative side, anyway experimenting with consciousness is a good thing.
They seem to be only using it in the background - "helping journalists interrogate large data sets, assisting colleagues through corrections or suggestions, creating ideas for marketing campaigns, or reducing the bureaucracy of time-consuming business processes"
If and when they use editorially, the seem to be checking and moderating with human oversight and "will be open with readers" when they do so