The impact of AI-generated content on journalism so far
The UK media industry is generating plenty of think pieces on the potential impacts of artificial intelligence and news on the changes it is already making. Going beyond the headlines, how concerned are journalists in reality about AI when it comes to their own work? Will the adoption of technologies like ChatGPT and Bard ultimately be a positive or negative innovation for journalism?
Our Journalist Voices by Vuelio panel considering the implications of AI included Press Gazette associate editor and New Statesman media correspondent William Turvill, Helena Pozniak, an independent journalist writing for the Telegraph, The Guardian, the Institute of Engineering and Technology and more, and freelance writer Amelia Tait, who contributes to outlets including The Guardian, The New York Times, Wired, the New Statesman, and VICE.
Already affected in their work by AI, the panellists discussed the possible problem areas ahead alongside the opportunities, as well as what PRs need to know about AI assistance in the creative industries.
It’s still early days for AI
As pointed out by William, even ChatGPT and Bard would admit that their technology is not 100% reliable and fool-proof just yet, and each of the panellists had examples of AI going wrong.
‘I was pitching an article, and I used ChatGPT for fact finding,’ shared Helena, who regularly writes about the impact of technology on society – ‘clean energy, to freedom of information, smart motorways to the environment’. The accuracy of information and reliability of sources is of utmost importance in her work – how did the AI app perform?
‘I’m so glad I double-checked the information it offered, because it had completely fabricated a massive landslide that killed thousands that never happened. So, I’m very wary and just playing around with it at the moment.’
William, who reports on the inner workings of the media itself, pointed out problems with bias already creeping into AI:
‘I’ve messed around with it. I asked it to provide a summary of the day’s news for me, and it wasn’t too good. When asked which UK news sources I could trust, it was very pro The Guardian and the BBC, but told me I couldn’t trust the Mail, the Mirror or The Sun. But I feel there is potential there’.
For Amelia, its use as an alternative to Thesaurus.com when searching for the right word came with feelings of uncertainty –
‘I asked ChatGPT to rework a sentence for me; I ultimately didn’t use what it suggested. It opened my mind a little more, but I felt a little bit dirty. I didn’t know what the ethics were on it’.
Helena offered that AI can already provide assistance on some elements of research for journalists – ‘It can summarise a research paper brilliantly and can do a lot of background research.’
What isn’t so great – the writing itself:
‘It’s just so bland. The copy AI apps come out with is so dire’.
Understanding the difference between content and journalism
‘I would distinguish between what is journalism and what is content,’ said Amelia.
‘I’ve worked for websites where you’re churning out content, and for that kind of thing, companies that aren’t investing much in talent could start using AI. And that comes with dangers on misinformation.’
That many journalists – and PRs – start their careers with duties that could be automated in future was a concern William spoke about:
‘It could be challenging for media companies that produce ‘clickbait’, or repurpose information from other sources. Those jobs are definitely at risk.’
‘Many journos don’t want to be doing that anyway, but there’s a danger of cutting off the entry level jobs into journalism; those jobs you have to do to find your bearings as a journalist. I would be concerned as someone entering the industry now.’
‘I’m sure it’s the same in PR – when you start out, you’re doing the unglamorous jobs. When doing work experience, I was walking a dog every day. AI couldn’t do that, but it could do the background research for a law firm. The ‘bottom rung’ could be in a difficult position’.
For Helena, the negatives would also reach audiences: ‘There might be a diminishing desire for longreads. You can see it on websites already with short-form summaries at the top. When time-pressed, are people really going to read something you’ve slaved over for days?’
Quality journalism requires human journalists (and journalists need human sources)
While coverage of AI can come with fearmongering, it is already embedded in parts of the journalist job successfully – as pointed out by Amelia, journalists regularly use AI transcription services for interviews, cutting hours out of the work of a writer:
‘We need to perceive these things as tools that we’ll use, that can help us rather than replace us.
I could waste ten minutes thinking of a particular word, and that’s not a skill or talent, that’s just time consuming. Using AI as a tool, that’s really encouraging and exciting’.
William underlined the importance of the human aspect of journalism. Ultimately, journalism has a human audience interested in human stories, and who better to share that than fellow humans (with assistance from AI on the admin side):
‘This has really solidified for me which journalism is going to be important in future as AI takes on some of the more basic writing and research – the journalism that journalists are going to want to do is original journalism. We will be looking for more personalisation, more research, more insightful interviews from PRs and a lot of thought going into pitches.
Something I’ve really been thinking about is stories I should be writing, I’ve set myself a test – could an AI do this research, if not now in five years. Is this useful? We’ll be looking for original stuff and any help with that is always appreciated.’
For more on how the media industry is covering AI, read our Vuelio Insights Report ‘Media attitudes to AI journalism’.
Want to connect with human journalists with stories of interest to their human readers? Check out how you can help them with sourcing experts, spokespeople, case studies and data via the Journalist Enquiry Service and find journalists covering your specialist topics on the Vuelio Media Database.
When I first started writing about logistics in the early noughties, the warehousing sector was dominated by the forklift truck and racking manufacturers. The majority of operations were performed manually – automated fulfilment systems were only really aimed at high-volume retailers. The advent of ecommerce changed everything, with the big warehouses that were previously configured to send pallets of goods to retail stores having to reshape their operations to satisfy the demands of individual consumers. This is where automation really took off, giving birth to a sub-sector called intralogistics – the technology of warehouse fulfilment. Believe me, there’s considerably more complexity required in a warehouse sending a single item to a single recipient than one sending a pallet of identical boxes to a supermarket.
