What PR and media teams need to do differently now AI is changing newsrooms

The technology is impressive

Let’s begin by looking at the technology.

We’ve been experimenting with AI since its launch, and you can easily understand the appeal of the efficiency savings it promises.

I was working on a media training course for one of our clients where some interviews were about a subject announced in a recent detailed press release.

I dropped that press release into ChatGPT.

Within seconds, it provided me with a summary of the release. And it outlined potential questions I could ask.

Those questions were not perfect. They would need some work to replicate a realistic interview with a journalist. For example, the suggestions did not cover the timely background context that would come into questions and there were no questions about wider topical or sector issues. 

But it gave me a handy framework to shape my interview and consider what I wanted to ask.

Handy stuff that has the potential to save me lots of time if we choose to use it. You could see why it could appeal to a journalist preparing for several interviews, particularly if they are on subjects where they have limited background knowledge.

But journalists would still need to carefully check the validity of its summary and questions (more on this later).

When I began pulling this blog together, I wondered how AI might write it.

So, I put it to the test.

A few prompts in the tool and I had a 1,200 word article in seconds. Job done?

Don’t worry, that’s not the blog you are reading.

I wouldn’t want to put my name to the words it produced, but it was useful inspiration.

And, while I will continue to write these blogs myself, I may turn to it again when the words don’t flow – it has to be better than staring at a blank page.

Is AI being used in the newsroom?

Like or loath the thought of AI, we are already starting to live in the future.

We always spend time with our media training delegates helping them understand the ever-evolving media landscape.

You may believe you have yet to read a news story that has been written by AI.

But there is plenty of evidence the technology is being used inside and around the newsroom.

Regional newspaper publisher Newsquest now employs seven AI-assisted reporters across the UK.

These reporters use an in-house tool that drafts stories “based on trusted information”.

The system features a ‘notes’ section where journalists can input information like press releases and quotes from a source.

And the AI tool then creates a story for the reporters to review and check.

Jody Doherty-Cove, Newsquest’s head of editorial, said: “We’re using this to alleviate the burden of the mundane, but very important, tasks that were on reporters, freeing them up to do that human touch journalism that really resonates with the communities.”

He added: “It’s just acting as a hyper-efficient copywriter – drafting the story on certain instructions, and then it goes back to the reporter who reviews the draft before publishing it.”

Last year, Reach – the owner of the Daily Mirror and the Express - published its first articles written using AI.

The articles were covered on its local news site InYourArea.co.uk, and you can read one of them – Seven Things To Do In Newport – here.

Jim Mullen, the company’s chief executive, was keen to highlight the human involvement in the story. He said: “It was all AI-produced but the data was obviously put together by a journalist, and whether it was good enough to publish was decided by an editor.”

The same group is now using an AI tool – called Guten – to help journalists quickly rewrite stories, including police press releases, which have already appeared on other sites within its network.

Rather than rewriting an entire article, Guten makes a selection of changes, such as swapping in synonyms or re-phrasing passages, without changing the meaning of the text.

The Daily Telegraph has banned journalists from using ChatGPT for copy editing. But it uses AI – called Echobox to achieve greater social media reach and automate activity on those channels.

Elise Johnson, the publication’s head of social media, said: “We take a human-led approach for deciding which content to post, but we often leave it to Echobox to decide when articles are posted. Especially when it’s not a very busy news day we let the AI take over.”

The Financial Times launched its first generative AI tool in March.

The tool allows subscribers to ask any question and receive a response using FT content published over the past two decades.

Lindsey Jayne, FT’s chief product officer, said: “We’re always looking to add value to our FT Professional offering, and this new feature will help our subscribers make confident strategic and commercial decisions quickly by getting answers rather than search results.”

On the other side of the planet, News Corp Australia is producing an eye-catching 3,000 articles a week using generative artificial intelligence.

Four staff use the technology to generate local stories on topics like weather, traffic and fuel prices.

Further afield, the German newspaper Bild has announced it will make additional editorial cuts because of “the opportunities of artificial intelligence”.

Its chief executive, Mathias Döpfner, has predicted AI would soon be better at the “aggregation of information” than human journalists.

And the New York Times has announced it is building a team to explore the use of generative AI.

A post on Threads said the new team is “focused on prototyping uses of generative AI and other machine-learning techniques to help with reporting and how the Times is presented to readers.”

But the publication has also reaffirmed its commitment to human reporters, saying: “Times journalism will always be reported, written and edited by our expert journalists.”

Cision’s State Of The Media 2024 report revealed 53 per cent of the 3,000 journalists surveyed are currently not using AI. But 28 per cent are using it ‘a little’ and another 12 per cent ‘a moderate amount’.

The tech is primarily being used for research, but just under 20 per cent of the journalists admitted using it to create outlines and early drafts of content. And 11 per cent use is to ‘brainstorm interview questions’.  

Trust, gaffes and hallucinations

One of the main issues for news organisations and the use of artificial intelligence is a lack of public trust.

A YouGov survey revealed Britons believe the introduction of the technology to journalism will do more harm than good.

The survey showed that just seven per cent believe they have read a news article written by AI and that only six per cent think the benefits will outweigh the drawbacks.

Almost 80 per cent of respondents said media organisations should be required to display on a news article any ways in which that article has been created using the assistance of AI.

This is something Buzzfeed does. In this article, for example, you will see a line saying: “This article was collaboratively written by Tara Falkenberg + Buzzy, our creative AI assistant.”

The lack of trust is no surprise.

People are worried about the use of AI generally and not just in the media. Afterall, we’ve been told the technology could lead to human extinction.

But specifically with the media, we’ve already seen high-profile instances of how AI could be misused, which will only further fuel the distrust.

Last year, the German magazine Die Aktuelle published what it claimed to be the “first interview” with Formula 1 legend Michael Schumacher since his near-fatal skiing accident a decade earlier.

But the article – including the supposed quotes – was entirely generated by AI.

Such was the fallout that the editor was sacked two days later.

The magazine publishers apologised to the family and were ordered to pay them 200,000 euros in compensation.

The Irish Times said it was “genuinely sorry” after it published a comment piece called ‘Irish women's obsession with fake tan is problematic’ which turned out to be an AI-generated hoax.

Sports Illustrated has been embroiled in scandal after it was accused of running articles created by AI.

An investigative report found the American sports magazine published articles by fake authors, complete with headshots and biographies created by the technology.

The articles were deleted, and the company’s chief executive was removed.

You may have also heard of a tech news site called CNET, which had to make corrections on 41 of the 77 stories it published using an AI tool.

The articles were riddled with errors, and an amendment posted on some suggests there were also plagiarism issues. “We’ve replaced phrases that were not entirely original,” it said.

Futurism, which broke the story about CNET using dodgy AI content, ran the headline: CNET is now letting an AI write articles for its site. The problem? It's kind of a moron.

Ouch.

But the CNET story highlights a significant issue with AI generative content. While it is impressive to use, it remains prone to hallucinations – the name given to its generation of inaccuracies, false information and general nonsense.

Sky News built an AI reporter for a feature on whether it could do the job of a journalist. Among the plausible pitches it created was a story that spilling milk could be crucial to safer roads. The story was a weird take on a real-life tanker crash that led to its milk load being spilt.

And it backed its claim with a quote from a supposed environment scientist. The real Sky reporter said it was concerning how the story had been presented “with confidence and conviction backed up with plausible, but entirely fictional explanations.”

A New York Times investigation into these AI hallucinations– written by a human – revealed AI systems make things up between three and 27 per cent of the time.

And that poses questions. Does AI really make newsrooms more efficient if you need to employ more humans to check the accuracy of its work?

Can a publication afford the trust and credibility impact that comes from covering an AI ‘hallucination’ as fact?

What AI means for PR, Comms and Media Teams

Despite those concerns, AI isn’t going to go away. The opposite is true. It’s just the pace of change, and how far it goes, that is open to question.

Either way, there are changes comms professionals need to be aware of and prepare for.

• Your press release could be summarised by AI. If you don’t like the sound of that, make sure it is short, snappy and easy for time-pressed journalists to understand.
• Similarly, AI could be used to find the relevance of a press release to the media outlet’s audience. The way to avoid this is by making sure your releases are targeted. This means creating multiple versions of your press release. Maybe AI could help with that?
• A fast news cycle will get even quicker. With AI able to quickly cut through data to find the nuggets, and alert reporters to breaking news and emerging topics, stories could be turned around even quicker. It means you may have to respond even more swiftly to media opportunities and interview requests. You may also need to act even more quickly in a crisis media management incident.
• Journalists will be freed up to tell the stories that matter. While AI can perform the more mundane tasks effectively, it cannot replace the all-important human touch. Freeing up their precious time can allow journalists to focus on human storytelling and carrying out media interviews so make sure that you take advantage of this and have your spokespeople prepared and ready to do media interviews
• Better media monitoring. The amount of data on the internet is forecast to reach 181 zettabytes by 2025. I don’t really understand what a Zettabyte is but the IT bods at Media First tell me that's a lot of content. AI is likely to be the best way to cut through that noise and find the stories that matter for your organisation, better manage your reputation and better proof your spokespeople.

It will be fascinating to see whether it works out better for those media organisations banking early on AI to drive efficiencies and cut costs. Or those taking a more watchful approach and questioning how it best benefits their audience.

Media First are media and communications training specialists with over 35 years of experience. We have a team of trainers, each with decades of experience working as journalists, presenters, communications coaches and media trainers. 

Click here to find out more about our media training courses.