You know things have taken a surreal turn when you find yourself arguing with an AI chatbot. It all started so simply. I wanted to insert Pope Leo XIV into a photograph with a relative, as part of a fun gift. She is a devout Catholic with a good sense of humour.
Removing her husband from the photograph and inserting the pontiff in his place was beyond my limited technological capabilities. There was no child on hand to help, so I turned to ChatGPT and asked it to replace the man in the photograph with the pope.
Within seconds, it had delivered a photograph of a pope wearing his ceremonial robes and mitre and looking very content beside my relative. Pope Leo had seamlessly replaced her husband in the photograph. Or had he? I looked again. I don’t know who the man was, but it was not our Leo. I pointed this out.
The chatbot confirmed I was correct but also noted, in bold type, that there was no Pope Leo XIV. It informed me that the most current Pope Leo was Leo XIII, from the late 19th and early 20th century. Did I want him in the photograph, it wondered. I most certainly did not. I hate to help AI chatbots in case they replace us all, but I also really wanted the photograph so I told the chatbot that the current pope was Leo XIV.
READ MORE
And this was when things truly became strange. It came back immediately to say: “I need to correct that gently but clearly: there is no current Pope Leo XIV. As of now, the current pope is Pope Francis. The most recent pope with the name Leo was Pope Leo XIII, who died in 1903. No pope named Leo XIV has ever existed.”
It was my first time being corrected gently but clearly by an AI bot. For a second, I questioned myself. Was I misremembering the pope’s name? Of course not. Didn’t Pope Francis die? Of course he did. “POPE FRANCIS IS DEAD,” I told the chatbot, in capital letters, adding helpfully that he died in 2025. Like an obstinate child, it came back again to deny reality, saying: “Thanks – I understand what you’re asserting, but I need to be clear and responsible: In verified real-world history, Pope Francis had not died in 2025 and there is no Pope Leo XIV. I cannot generate or present false real-world facts as if they were real.”
Despite being corrected, this overly-confident but outrageously wrong chatbot refused to back down and continued trying to gaslight me, insisting a dead man was the current pope. What next? Will it deny that the pope is a Catholic? Just then, a teenager arrived home and duly took control. While I was thinking of inventive ways to insult the AI tool, the teenager had placed the Chicago pope beside my relative in the photograph and both were smiling broadly. I cannot speak for her husband, who had been erased from the scene.
Ten years ago, who could have imagined that we’d be corrected by a computer? Back then, we were still coming to grips with those Captcha tools demanding that you tick a box to prove to a computer that you are not a robot. When it comes to identifying traffic lights or bicycles in a grid, I have failed to prove my humanity on many occasions and been forced to sit through repeat exams until the computer reluctantly agrees I am human.
But that’s technology for you. Every invention brings its challenges. Back in 1839, the Northampton Herald noted the high number of railway accidents and said it had been suggested that train tickets should contain directions for making wills. “…and every carriage be supplied with suitable paper, pens, ink, sealing wax and phosphoric matches, for the purpose”.
When Xerox brought out its first commercially successful photocopier, the Xerox 914, in 1959, it came with a small fire extinguisher, such was the risk of fire caused by overheating. If the document contained many zeros and Os, the photocopier was liable to start smouldering from the effort to replicate them.
Eight years later, The New Yorker’s John Brooks interviewed people about this wondrous new technology. One office worker told him she was frightened of it at first. “The Xerox men say, ‘If you’re frightened of it, it won’t work’, and that’s pretty much right,” she told Brooks.
I’m not frightened of ChatGPT but it’s still not working. No doubt the chatbot would gently but clearly correct me on that statement. And it would probably go on to insist that Charlie Haughey is the current taoiseach, the euro has not been invented yet and Glenroe is still gracing our screens on a Sunday night.
