We are all acutely aware of the damage that AI-generated content is inflicting on our industry and craft.
And I’m sure I am not alone in despairing at the tsunami of ‘content’ some PR agencies are churning out that is light on evidence, heavy on product placement and always comes with a demand to use a boilerplate sign off with even more branding and a web address. Indeed, our current home-page poll found that 73 per cent of members have noticed an increase in rubbish press releases.
But I recently heard two similar accounts of freelance journalists getting email responses to a #journorequest for case studies on social media which they suspect were AI generated.
Case studies can often make a story, but we all know they can be tricky to find — so we cast a wide net via platforms like ResponseSource, Mumsnet, X and Facebook.
Too good
But as the recent example of the Telegraph’s mea culpa for publishing a bogus case study shows – some case studies really are too good to be true.
Accounts vary, with some claiming the Telegraph story was AI generated, but the Press Gazette reported that the bylined journalist — a well respected freelancer who I shan’t name — did actually interview the pretender.
The warning is clear. Check, double-check and triple-check any case study sourced in this way exists. We all leave a digital footprint, if your potential case study hasn’t that should be an immediate red flag.
And if a case study says they can only answer questions by email, ask yourself why? Having had an email response from what seemed the perfect case study, a freelance friend told me she became increasingly suspicious when the would-be interviewee simply refused to be interviewed.
Their email replies provided various excuses, and reassurances that their story was genuine — but all communication stopped when the freelancer insisted on speaking to them.
Due diligence has always been an important part of our job, and with AI and the proliferation of digital PR agencies and marketeers it is more important than ever more. Writer, beware.
Recent Comments