In February 1887, just as the foundations were being dug, a group of France’s most famous artists and writers sent a furious letter to a Paris newspaper. Their target was Gustave Eiffel’s planned tower, which they said would make the city “irreparably ugly”. It opened two years later anyway. One of the protesters, the writer Guy de Maupassant, then took to eating lunch in the finished tower’s restaurant. According to the philosopher Roland Barthes, his reasoning was simple: it was the only place in Paris where he couldn’t see the thing. Today you can’t sell a postcard of Paris without it. Psychologists call this the mere exposure effect, and some call it the Eiffel Tower effect. In 1968 Robert Zajonc showed people made-up words, some once and some up to 25 times, then asked them to guess which ones meant something good. The more often people had seen a nonsense word, the nicer they assumed it was. Nothing had changed except how familiar it felt. I thought of Maupassant this week while reading a new Gallup survey of how people in 37 countries feel about AI. Most of the world is fairly upbeat about it. In 34 of those 37 countries, people have more good feelings than bad about AI, and the most common feeling of all is curiosity. The big exception is the rich West. Americans are the most worried people in the whole survey, with three in four saying AI worries them. Brits aren’t far behind, at almost two in three. Meanwhile Nigerians are nearly twice as likely as Americans to think AI will be good for their country, even though four out of five Nigerians have never used it. Within those worried Western countries, the calmest people are the ones using AI every day. In the US, daily users are more than five times as likely as non-users to trust what AI tells them. Gallup can’t say which way round this works. Perhaps trusting people simply use it more. Zajonc would have bet on familiarity doing a fair bit of the work. Why this matters if you’re trying to change minds Most attempts to get a team using AI rely on persuasion. A town hall, a deck of productivity stats, a stirring email from the CEO. That’s roughly the equivalent of writing to Maupassant to explain the engineering merits of wrought iron. Familiarity can’t be argued into existence. People have to have lunch in the tower. There’s a second lesson for marketers. Gallup found that almost everywhere, people are more hopeful about AI than trusting of it. Most AI marketing pitches straight into that gap with fear. “Don’t get left behind.” “Future-proof your career.” Those lines sell anxiety to an audience that has plenty already, and anxious people put things off. Curiosity is the emotion with the biggest audience, and it’s the one that gets people to try something. Trying leads to familiarity, and familiarity is where the trust comes from. Paris didn’t come round to the Eiffel Tower because somebody won the argument. It came round because the tower stayed put and people kept walking past it. One Key Takeaway Familiarity does what argument can’t. If you want people to feel better about AI, get them using it in small, low-stakes ways, and sell the curiosity rather than the fear. How to Use It Swap the town hall for a tiny daily habit. Ask your team to hand their AI tool one dull ten-minute job each morning for a fortnight, such as summarising a long email thread. Liking rises fastest over the first few exposures, so the first week does the heavy lifting. Lead with a question. In campaigns and internal comms, replace “Don’t get left behind” with something people want to poke at, like “What would you ask it first?” Give the sceptics a sandbox. Your most worried colleagues need a safe place to play and a real task to do, far more than they need a better argument. The worry tends to shrink once the thing is familiar. Want to go deeper? Our AI for Marketers course is a practical way to build exactly that habit on real work. And if you’d like to know why familiarity, curiosity and trust do so much of the heavy lifting in persuasion, Behavioural Science for Brands is full of it.