There has been much talk about the quality of AI-written emails — experts say they can spot one a mile off.
But what about the public? Can they tell the difference, and do they even care?
Instiller, a UK company, recently sought to answer that question. It used ChatGPT to turn out three emails for a fake beauty brand, and assigned a professional copywriter to do the same.
Only 1.18% of consumers were able to correctly guess the origin of all six emails. Worse yet, 52% incorrectly guessed that the copywriter’s emails were AI-written, and 48% said vice versa.
The final score: Half overall were unable to guess the difference between AI and human-written copy. Men were more likely than women to spot an AI-generated email at a margin of 51% to 48%.
On an age basis, Gen Zers were the most accurate. People in the upper age brackets did worse, although only marginally so.
Here are the age groups and the percentages of correct responses:
In general, 73% were unconcerned with the idea of AI-generated marketing campaigns. Oddly, only 20% in the 55-64 age cohort were upset about it. But 30% of those over age 60 were troubled.
Gen Z, the sharpest at spotting AI-written emails, are the most worried about it—37.5%.
It’s not clear how Americans would score on this test, but we guess they’d be roughly equal to their UK counterparts.
Instiller’s conclusion? “This experiment has real implications for the future of digital marketing,” it writes. “Clearly, the public is torn, with the results being completely split. That said, it’s clear the distinction between AI-generated content and content created by humans is becoming more and more indiscernible, and less and less people care either way, which makes AI a useful addition to any email marketers arsenal.
We hope they had proper consents for sending these emails, and that nobody complained about fraud.
Instiller surveyed 510 consumers from across the UK.