Unrecognizability Is the Newest Design Principle

What should a space offer people when they don’t want to be recognized at all?

Most people don’t object when a hotel remembers they asked for a quieter room, a workplace saves a lighting preference, or a building makes access easier because it already knows they belong there. Personalization works when it removes friction and gives something useful back. The problem starts when the same systems move from remembering what someone asked for to deciding who they are, what they want, whether they belong, or what they’re likely to do next. Buildings, products, and services are getting much better at recognizing faces, voices, movement, devices, habits, purchases, and patterns of behavior, and the selling point is usually convenience. The room knows what you need before you ask. The store anticipates what you want. The building adjusts itself around you. It all sounds wonderful until being recognized starts to feel like a requirement, and people begin asking for spaces, experiences, and products where it isn’t.

Parts of the European Union’s AI Act are beginning to show where this could go next. The act prohibits several uses of AI recognition, including emotion recognition in workplaces and schools, certain forms of biometric categorization, and most real-time remote biometric identification in public spaces, with limited exceptions. To me, that points to a new area of advocacy and design centered on when we can and cannot be recognized in a space, and designers are already beginning to explore options beyond simply turning the technology off or checking a box in an app.

Eric Lu’s Ghost Font is one of the latest examples of design tackling unrecognizability. It hides letters inside fields of moving dots. People can still read the word because the brain groups the motion together, while current AI systems often struggle to identify it. It took less than a day for an AI to learn to read it, but that’s almost beside the point. The font treats machine unreadability as a design feature.

Researchers are doing something similar with materials. In July 2026, a University of Illinois team reported building the first three-dimensional thermal cloak that can hide an object from infrared detection no matter which direction the heat comes from. Hyperstealth’s Quantum Stealth material takes a different approach by bending light around a person or object, creating the appearance that parts of it have disappeared into the background. It doesn’t make something truly invisible, and the effect depends on viewing conditions, but it shows that materials can be designed to interfere with how both people and sensing systems perceive what’s in front of them. These aren’t Harry Potter cloaks, but they suggest that unrecognizability may eventually be shaped through materials, geometry, and the surfaces around us.

Anti-surveillance clothing, reflective surfaces, face coverings, and makeup designed to confuse computer vision are exploring the same idea. Some work better than others, and the systems will keep changing, but together they point to something larger. Unrecognizability is becoming designable, wearable, applicable, and its own form of personalization.

Now, a design principle earns that name when it changes the decisions we make. Unrecognizability would affect where cameras are placed, how access is handled, which materials are specified, how circulation is organized, what information a room is allowed to keep, and whether people can use a space without first becoming a profile. That’s why I think this is moving beyond a privacy preference or a technology workaround. It’s becoming a new way to decide what a space should and shouldn’t be able to know. That is a new type of conversation that should be had more.

I could see the lobby becoming the first place this shows up. It already concentrates cameras, access control, visitor systems, license plate readers, QR codes, and security screening. In the future, it may also need to show where recognition ends. A person could verify that they’re allowed into the building without being continuously identified everywhere they go afterward, or choose a low-data route through the space. If this scales into entertainment venues, airports, or other high-traffic places, those circulation decisions could get complicated very quickly. The places where we eat and gather socially may be next. Seating areas could offer cloaked and uncloaked zones, creating a new kind of choice around privacy and security. A retail space might let someone browse without linking every pause to a profile, while a workplace café could count occupancy without recording who sat where and for how long.

This doesn’t require every room to become an underground bunker lined with anti-camera wallpaper. Lighting could limit certain kinds of capture, partitions and layouts could interrupt continuous tracking, and materials or graphics could make machine vision less reliable. Signs might show where sensing is active, while manual controls could remain available for people who don’t want the room to personalize itself around them. Some of this will be technically difficult, and some of it will create conflicts with security, accessibility, operations, and safety. Unrecognizability won’t mean that every person is invisible everywhere, nor should it. There are places where identification is necessary. The question is whether recognition needs to continue across the entire experience simply because the technology can do it.

For the field, this may become a real differentiator. Owners may start competing on how much control people have over recognition. Developers could use low-data spaces as a leasing or hospitality advantage. Manufacturers may create materials, finishes, lighting, and screening products that interrupt machine vision or make sensing boundaries visible. Technology providers will need clearer off modes, local controls, and systems that verify access without tracking people everywhere afterward. Operators will need to prove that an AI-off room is actually off, while insurers, lawyers, and procurement teams will have to decide what counts as a credible promise.

We've already designed for visual and acoustic privacy, secure thresholds, screened entrances, and places where people can step away from one another. Designing the right to be unrecognized may be the next version of that work, and one that is wide open to be led.