Help! I nearly fell for an AI scam

They say that everyone could fall for a scam at least once. I suspect this is true. It happened to me once. I still think about it sometimes. I was six years old, maybe seven, and walked to the front door to check the mail. What I found blew my mind. Zinedine Zidane – the Zinedine Zidane, at the height of his powers in the late 1990s – was coming to our local square that weekend.

I remember my heart pounding as I ran to tell my mum about it. She read it, read it again, softly laughed and explained to me that it wasn’t true. Zinedine Zidane, arguably the single most famous football player in France at the time – maybe the world – wasn’t coming to our neighbourhood in Nantes. I was crestfallen.

Though I did eventually recover, I still wonder what the purpose of that leaflet was. My parents don’t remember it, as the incident was far too minor to register in their adult memory, so the mystery shall forever remain. In any case, the incident was a useful one, as it taught me, at a very early age, not to trust everything I read or saw.

I’ve not fallen for any other kind of scam since then, which feels like a minor miracle given how prevalent they eventually became online. Amazingly, it seems they’re now getting worse than ever, with AI tricking more and more people into spending money on non-existent things, or even going to fake places and events.

Most recently, hundreds of people turned up in Birmingham’s Centenary Square to watch New Year’s Eve fireworks displays that never came – for the second year in a row. Back in November, scores of tourists schlepped across London to Buckingham Palace, having seen pictures of what they thought was a delightful little Christmas market in front of the gates. There was, however, no such thing: all the images circulating around social media had been fabricated.

This has now become quite the global problem. Over in China, people apparently kept looking for an Eiffel Tower in Beijing; in Peru, hikers asked guides to lead them to sacred temples and valleys which only existed in their imagination – and on their screens. Though some scams are purposely spread across the internet by malevolent forces, many of the mistakes are, in a way, honest ones, and stem from people’s over-reliance on generative AI.

Because, say, planning a holiday can be a lot of work, an increasing number of people now rely on websites like ChatGPT to create itineraries for them. Clearly, they don’t know or care that AI technology is still significantly flawed, and likely to frequently hallucinate results.

This is, of course, bad news for people like me. I loathe genAI and all the companies pushing it to people now, and wish more users could see the error of their ways. To be clear, I’m not against the very idea of AI: there are obviously some terrific uses for it in quite narrow scientific fields, for example.

What I object to is this idea that you can simply release a half-finished product, blithely refuse to care if and when it makes mistakes, and not accept any responsibility when things go wrong. See, for example, the horrors currently unfolding on Grok. There is a world in which I could have been an enthusiastic AI advocate, but it simply isn’t this one.