A robot has “fallen at the first hurdle” after being challenged by a data journalist to do her job.
Aimee Stanton, who works for National World’s data team, has deconstructed a “poor piece of journalism” by ChatGPT after challenging the AI tool to “write an article about NHS A&E waiting times in England.”
In a feature about the experiment, Aimee revealed the robot had “fallen dramatically at the first hurdle in becoming a journalist” because it failed to use up-to-date NHS statistics as a source for the artificially-created story.
However despite the hiccup, Aimee urged fellow journalists that they “need to get with AI and use it to our advantage”.
Said Amy: “My AI assistant claimed ‘over 300,000 patients had to wait longer than four hours to receive the care they needed’. While this claim is technically correct (483,000 people were waiting over four hours in January and 551,000 in March), I assume it hasn’t just gone with some obscure rounding technique and is just publishing wrong information.
“Making up figures is generally seen as a sacking offence at National World but let’s see if it improves.”
“When I first pitched this story idea I’d be lying if I said I wasn’t a bit apprehensive about talking myself out of a job but needless to say I’ll still be working as a journalist for a while yet.
“AI is certainly not to be undervalued by the journalism industry though. We would be utterly foolish to not see what is staring us straight in the face and the benefits/risks that come with it.
“Whether AI will replace me is up for debate and prior to asking ChatGPT to write me an article on A&E waiting times, I asked if it would.
“It said what we all already know: ‘Journalism requires a level of human intuition, creativity, and critical thinking that cannot be replicated by machines.’
“Thanks for the reassurance, ChatGPT, but I’m not so confident AI won’t radically affect the industry.”
Aimee’s experiment comes after Reach plc revealed it has begun testing the ability of chatbots to aid its coverage.