Philosophy
Data Humanism, my manifesto for a new data world, 2017
Data is the most powerful force in society today. Data is everywhere—present in every moment, every event, every transaction or interaction with someone else. Every time you send a text, call a friend, fill out a form, hail a taxi, stream a movie, surf the web, pay a bill, buy groceries, buy anything, take your temperature, count your steps, swipe right (or left), you generate data. There’s data in the weather, in the air, in the ground, in outer space. If you own a smartwatch, you carry data on your body. If you have a cardiac pacemaker, you carry data in your body.
We are surrounded by an invisible ecosystem of data, and it’s becoming harder and harder to ignore.
From the rise of artificial intelligence to the specter of global climate change, today data makes and shapes our lives. But what is data, really? It’s a question that is surprisingly hard to answer. To some, data means numbers: figures on a screen, dots on a graph. It’s also often (falsely) equated with facts, an invariable form of concrete knowledge that always tells the truth. But in reality, data is hardly so incontrovertible. Data is an abstraction of reality, a useful but imperfect representation of real life.
Like life, it’s full of nuance, imprecision, and ambivalence. It’s quantitative and it’s qualitative. And it’s made by us—humans. No surprise, then, that data is often riddled with human bias.
My obsession with data began in childhood, counting and sorting the contents of my grandmother’s tailor shop in Italy. Data helped bring logic to chaos, even if the nebulous taxonomies I used to organize the shop’s buttons and ribbons were only in my head. In college, I discovered architecture, an entire profession rooted in numbers, and became infatuated with urban mapping: how one could visualize, through data, the spatial and social footprint of a city. The more sophisticated your map became—the more data it took into account—the more true to life it was.
Today, my team at Pentagram helps companies and individuals communicate data to the wider world. While most designers use color, shape, or typography to express ideas, we use data. Sometimes this looks like a traditional graph or a chart, but more often it takes unconventional forms: a sculpture, a website, a mural, even an article of clothing. Really, anyone can use data as a lens through which to understand the world. You just need to know how. Akin to a pair of glasses that can be taken on and off, data is a filter to find pattern, structure, and ultimately meaning, one subject at a time. I call this superpower Data Humanism, and it’s a radically different way to think about, understand, and engage with numbers. As the name suggests, Data Humanism centers people, rather than numbers, in its conception of data. Just as the humanist movement during the Renaissance sought to emphasize individuals rather than God as the ultimate source of meaning, Data Humanism underscores the foundational role of people—us—in data’s collection, analysis, and communication.
Despite how it may sometimes appear, data is always the product of human hands.
In our hyper-complex, ever-changing world, I don’t believe there’s just one definition of data. There can’t be. Data’s pervasiveness means that there are multiple definitions—multiple meanings—that must be contended with if we are to truly understand the full potential of data in our lives.
As data becomes further entrenched in society, all of us will need to think more critically about the data we produce, consume, interpret, and manipulate daily.
This is the same argument I have been making for ten years, since I wrote the Data Humanism manifesto.
With A.I. the conversation is no longer about what data is, but about who does the choosing, and the investigating: understanding what data can say, and what it means, has to stay in our hands, also when we use the tools A.I. gives us.
Ten years ago, for one whole year, I collected data about my own life by hand, and so did a friend, another designer, on the other side of the ocean. Every week we each chose one small thing to pay attention to and count, some ordinary part of our days, and we turned that week into a hand-drawn drawing on a postcard, and mailed it across the Atlantic. Every dataset is made of choices about what somebody thought mattered enough to write down. When I made mine, I made those choices, one week at a time. When the machine does it, the choices are all still in there. They have been designed out of view.
With A.I., the authorship around data gets even more hidden. The answer arrives clean and finished and authorless, in a confident voice, and we do not even consider what we get back from a chatbot as data anymore. But that is data. It is the output of an enormous training set, full of choices about what got collected and what got left out, all hidden in a black box, and handed back as if it were certain. And because the system is built to give back the most statistically likely answer, it smooths everything toward the middle.
Human experience is complex and particular. Creative work means contemplating and translating that complexity rather than flattening it into the simplest answer. This is true whether you are making a data visualization, a brand, a story. When we reach for the wrong tools too quickly, we outsource the part of the process where our particular experience and point of view would have entered the work.
In my own practice the process looks something like this. You talk to people. A lot of them. You listen and you learn about their needs and specific context. You dig through visual references and accidentally find something you were not looking for. You push things around on paper, or on a screen, and land somewhere unexpected. You fail in a way that opens up a completely unexpected direction. You revisit. You refine. You look in another direction. Refine again. And the wandering is where a lot of the actual design happens.
My team does use A.I. to help clients implement the designs we’ve created at scale. That is A.I. doing what it’s good at: taking something already defined by a human and iterating on it efficiently. Where we won’t use A.I. is during the discovery, the listening and thinking part.
A few months ago, I started making things for no reason: painting, sewing paper patterns, scribbling on rice paper with illegible handwriting. This practice has begun to inform everything else I do: the way I look at references, the way I sketch, the way I approach a project before opening any software. If we are people who make things, and I take real growth and pleasure in the act of making, who are we becoming if we do not make anymore? I do not worry that we will stop making things. I wonder what happens if we stop being made by the things we make.
Anything worth remembering came from a person who did not know where they were going until they arrived.
— Giorgia Lupi, 2026
In Defense of the Detour, The New York Times, September 2026
Data Humanism, my manifesto for a new data world, 2017
Watch my TED Talks to Learn More
How we can find ourselves in data
TEDNYC, March 2017