My childhood was not especially remarkable. I was the youngest of three kids, with two older sisters who were, for all practical purposes, my bosses. They made me play school, listen to their records, watch their after-school shows like it was a religion and regularly conscripted me into their barely rehearsed puppet productions. In our basement stood a small white wooden puppet theatre with an old printed quilt curtain, all of it surrounded by the particular grandeur of 1980s wood paneling. My sisters took these productions seriously, me not so much. There were acts, characters, storylines, costumes, intermissions, snacks. We were no Von Trapp family, but we tried our damnedest. I was rarely the headliner – instead I was cast as the family pet, the unnamed child ‘Guard #2’. Real bangers.
My parents would come downstairs, we would dim the lights and about halfway through the show I would inevitably get bored. I would look at whatever puppet was on my hand and decide it was time for chaos. I would burst into scenes where I did not belong, eat the other puppets, knock over the scenery, maybe throw a few elbows. The whole thing usually ended in a fight. But, like any committed method actor, I never broke character. It did not really matter whether the puppet was a tiger or a knight or some floppy piece of felt with two eyes glued to it. Once it was in my hand, it was alive. It had motives. Opinions. A laundry list of enemies. To become somebody, all it took was a little felt, some glue and a hefty dose of make-believe.
Lately I have been thinking about that basement a lot, because suddenly everyone in AI seems to be making puppets. Meta has Muse. OpenAI has Dots. Anthropic has Claw’d. Even Notion has its little face. Some of the most advanced technology companies in the world are building systems capable of searching, reasoning, remembering and acting on our behalf, then surrounding those systems with names, faces, personalities and little bodies. Which raises a question I can’t quite shake: why does the future need so many little guys?
You could write the whole thing off as branding. Mascots work. We know this because we grew up on a steady diet of them: Tony the Tiger, the Michelin Man, the Pillsbury Doughboy, Mr. Clean and Grimace. And mascots are everywhere again well beyond AI. Sports mascots have become celebrities. Old characters are being pulled out of dusty archives. Brands that once spoke through immaculate typography and white sets are letting weird little creatures loose on social media. But nostalgia alone does not feel like a satisfying explanation. Something else seems to be going on. Simply: the world is weird right now.
We left quarantine and got thrust immediately into technological change. We are connected to everybody and somehow still talk constantly about loneliness. We are surrounded by systems that are incredibly powerful and impossible to understand. The people I work with feel it. Clients feel it. Friends feel it. There is excitement but also anxiety and uncertainty. The ground is jello. And when the ground turns to jello, apparently one of the things humans do is put a face on it.
A while ago, while trying to understand why, I came across an academic paper called On Seeing Human by Nicholas Epley, Adam Waytz and John Cacioppo. The paper is about anthropomorphism, our tendency to give human motivations, personalities and emotions to things that are not human.
Why does a car look friendly? Why is the computer being stubborn? Why does a storm feel angry? The framework is academic, but the underlying ideas are pretty simple.They even developed a handy acronym to explain what they found. It’s called ‘Seek’, and it stands for sociality, effectance and elicited knowledge.
Sociality is our desire for connection. We are social creatures and, when connection feels scarce, we start finding it in other places. We talk to pets. We name our cars. We yell at the stop light. Effectance is our desire to understand and predict the world around us. When something is complicated or unfamiliar, we reach for explanations that make it legible. Your computer freezes and you say, “It’s being stubborn”, rather than launching into a diagnosis of memory allocation, CPU cycles and thermal throttling. “Stubborn” works just fine. Elicited knowledge is the tendency to use what we know best (ourselves) as the model for understanding something that isn’t us. Two headlights and a car grille become eyes and a mouth and suddenly the car looks happy. A circle with eyebrows looks worried. A blinking dot looks like it’s thinking.
Put those things together and the current AI mascot explosion starts to make a certain kind of sense. We want connection. We are confronting technologies most of us do not completely understand. And those technologies increasingly behave in ways that remind us of ourselves.
We talk to them and they talk back. They search for us, remember things, recommend things and increasingly take actions on our behalf. Yet what is actually happening under the hood remains almost completely incomprehensible to most people. AI, in other words, presses on all of the buttons that make us anthropomorphise. The more alien the technology becomes, the more urgently we seem to make it familiar.
That’s what I find so interesting about Muse, Dots, Claw’d and the rest of this emerging cast of characters. None of these systems needs a face in order to function, but we may need the face in order to understand how to relate to the system. There is something almost comical about it. We have built technology that can feel impossibly sophisticated and abstract, then immediately find ourselves reaching for one of humanity’s oldest tricks: give it eyes, give it a name, give it a personality. Suddenly we have a way in.
I have had the chance to see that impulse from the inside and believe it really is the best way to adapt and adopt. Buck partnered with Notion on its AI experience, working with a face originally designed by the illustrator Roman Muradov. Roman had done something incredibly economical: a few lines, eyes, brows, a nose. Our job was to breathe life into it. We did what animators do. We began asking questions. How does this thing think? What does it do while it searches? How does it greet you? What does waiting feel like? What does an error look like? What happens when it’s writing? The goal was to take Roman’s beautifully simple face and turn it into something that could behave.
That’s where animation becomes a kind of magic. Give something timing and suddenly it appears to have a thought. Give it anticipation and it seems to have intention. Make it hesitate and now it has doubt. A few pixels can become shy. A line can become angry. A circle can become curious. As we developed the Notion character, those behaviours could be plugged into different states – thinking, writing, searching, greeting, error – and extended over time. The face stayed simple, but the character began to feel alive. We were not trying to make the AI itself more impressive. Notion already had the technology. We were helping to make an unfamiliar interaction warmer and more understandable, to distill and humanise.
That, to me, is mascot 101, and it may also be why so many people have responded to the Notion character. It does not explain the technology, rather it gives us an emotional foothold into it. This distinction matters, because there is a tension inside anthropomorphism that’s especially important with AI. A face can make something complicated easier to understand, but it can also make something complicated seem more understandable than it really is. Personality can create familiarity and familiarity can create trust – that trust still has to be earned. There is a difference between helping someone understand how to interact with a system and encouraging them to believe the system is more human, more knowledgeable, or more emotionally aware than it actually is.
I do not think that tension makes anthropomorphism bad. It makes it interesting. The urge to humanise these systems is telling us something about what we need from them and, perhaps more importantly, what we need from one another. When technology becomes difficult to understand, we do not necessarily retreat from humanity. We look for it. That’s why I am less interested in declaring the current mascot boom a trend than in what the trend reveals. We keep hearing that AI will make the future less human, while one of our most immediate responses to AI has been to surround it with signs of humanity. We give it names, faces, voices, moods, quirks, backstories. We make it cute. We make it strange. Sometimes we make it mildly threatening. Then we wait to see whether people will care.
The good ones give us something a logo or an abstract piece of software cannot: someone to have a relationship with. A character can surprise us, annoy us, disappoint us, get into trouble, develop a reputation, disappear for a while and come back with a new haircut. That’s why the best mascots are not really logos with arms. They are storytelling systems. And storytelling systems require actual storytelling. You need to know their personality, their backstory, their voice, their arc. You need to know what they will do and what they will never do. You need to understand what made them, even if what made them is just dough and sprinkles. Most importantly, you have to believe in them, because if you don’t, nobody else will.
For all the talk about AI changing creativity, I find it reassuring that we are still using such old creative tools to make sense of it. Character. Performance. Story. Make-believe. The output may be digital, the technology may be extraordinary, and the tools we use to make things will continue to change. But making something feel alive remains a craft. It takes observation, empathy, timing, humour and often restraint. It takes somebody caring enough to decide exactly how an eyebrow should move when a little face is thinking. It takes fingerprints.
Maybe that’s why this mascot explosion gives me a strange amount of hope. We are encountering machines that can do things that would have seemed impossible not very long ago, and somewhere inside all of that complexity humans keep doing what humans have always done. We tell stories. We make characters. We see faces in the clouds. We take something we do not quite understand and imagine that maybe, somehow, there is somebody in there.
Same as it ever was. We are still kids in the basement, telling stories to ourselves and to one another, using a little make-believe to make the world less scary and ourselves a little less alone. And, if we are doing it right, leaving just enough room to cause a little chaos.