---
title: "This printed bottle announces when its cap is loose — with no electronics inside"
locale: "en"
url: "https://irz.fr/en/articles/shiftlens-printed-state-in-material-en"
markdown_url: "https://irz.fr/en/articles/shiftlens-printed-state-in-material-en.md"
category: "tech"
tags: ["3D printing", "fabrication", "interaction", "optics", "MIT", "CSAIL"]
published_at: "2026-08-21T09:00:00.000Z"
author: "Arthur Lacoste"
translation: "https://irz.fr/fr/articles/shiftlens-etat-dans-la-matiere-fr.md"
---

# This printed bottle announces when its cap is loose — with no electronics inside

ShiftLens, a system from MIT, produces objects whose surface changes with the user's gesture: a bottle turns green when its cap is tightened. No sensor, no screen.

On a laboratory shelf sits a chemical bottle that looks like any other, molded by a 3D printer and carrying no screen at all. Turn its cap all the way and the bottle turns **green** with a check mark. Loosen it and the surface flips to **red**, the check becoming an exclamation mark. With that gesture alone the bottle tells a human being whether the container is properly closed or about to leak, without a sensor, without a battery, without a printed circuit, without a wire. The only source of the visible change is a mechanical gesture: the rotation of the cap.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

That object appears in the announcement video for **ShiftLens**, a design and fabrication system developed at MIT in the HCI Engineering group run by Stefanie Mueller, a member of CSAIL, in a paper led by Yunyi Zhu (MIT EECS) with Narjes Pourjafarian (Northeastern), presented at the UIST 2026 symposium.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)[2](https://hcie.csail.mit.edu/)

The opening promise fits in one sentence: let anyone without training in optics or mechanics fabricate an object whose surface **changes appearance with the interaction**, as if the material itself carried a display. The result is an object printed in a single pass on a multimaterial 3D printer whose visible state follows what the person holding it does.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The optical trick is not what matters most here. What matters is the decision to replace an entire electronic chain with **lenses, printed patterns and a mechanical shift**, so that sensor, microcontroller, screen and battery all disappear. The object is not easier to design; only the constraints it can withstand change completely.

## No screen required

Every interactive object in our daily life displays its state the same way, through a screen or an LED driven by electronics that measure something, and the habit is so standard that we forget how fragile it becomes outside the living room.

The MIT researchers list the conditions that kill electronics: water, aggressive chemicals, crushing, twisting, heavy pressure.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805) A laboratory bottle holding a corrosive solvent has no desire to carry a printed circuit, and neither does an outdoor warning sign, because the moment the display is electronic the environment of use becomes a permanent risk.

The classic alternative has existed for a long time: static labels, curved lenses that change appearance with the viewing angle. It only supports one kind of interaction, the movement of the viewer's gaze relative to the object, so the object does not react to a gesture at all.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805) It only changes according to *where you look at it*.

ShiftLens attacks precisely that limit. The object's appearance is no longer commanded by viewing angle but by **a mechanical movement, internal or external**: the rotation of a cap, the sliding of a drawer, the flip of a switch. The person controls the display directly through the gesture they make on the object.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

## Two thin layers

The core mechanism comes down to two stacked elements on the object's surface.

The first is a **lens layer**, an array of tiny lenticular lenses, each one a small cylindrical lens that deflects light differently depending on the angle you look at it. That is the principle behind 3D postcards and stickers that change image as you tilt them.

The second is a **backplane**, a layer of patterns: strips of images that correspond to the object's possible visual states. For the bottle there are two states, green plus check and red plus exclamation, so two sets of patterns printed side by side on the same backing surface.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

When the user shifts the lens layer relative to the backplane, **different parts of the pattern move under the lenses**, the lenses magnify them, and the surface shows one state or the other. The displacement can be tiny, fractions of a millimeter, and yet what the eye sees changes completely.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The technical difficulty sits exactly where you would expect it: making three things align that have no reason to do so spontaneously. Yunyi Zhu, the lead author, puts it plainly in the MIT announcement: “the biggest challenge in this project was to make sure all moving parts align: the optical effect, the mechanical linkages and the computational graphics”.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

## The gesture does the work

The trick of ShiftLens is that the displacement between the two layers comes naturally from the object itself, because no motor slides the lenses: the gesture of use produces the movement.

To see why that is clever, remember what a lenticular lens does. Those small cylindrical lenses deflect light differently depending on the arrival angle, and you find them in 3D postcards and in sleeves that change image when tilted. In a postcard the displacement comes from your hand tilting the card or your head moving in front of it, so the interaction stays involuntary in a strict sense: you did not ask to see the other image, you simply changed your point of view.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

ShiftLens reuses the same optical device but replaces the movement of the gaze with movement of the object itself. The lens layer is no longer rigidly attached to the backplane, so it can slide, rotate or travel relative to it, and this movement, produced by the gesture, brings up the other state.

There is an immediate practical consequence: the user does not need to be in a precise position to read the state, because whatever angle you look at the bottle from, you see the state selected by the gesture rather than the state selected by the geometry of your gaze. The display becomes deterministic from the user's point of view, since a given gesture always produces the same visible state whatever the position of your head.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The simplest case is an interaction that is already mechanical in the object **before** anyone thinks about it. A lipstick tube already rotates to push the stick out, a bottle cap already rotates to close, a switch already flips, and ShiftLens inserts itself into that existing movement: the rotation of the cap optically shifts the two layers relative to each other and the displayed state follows the gesture.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

When the object has no natural movement, you add one: a button, a slider, a roller. The researchers built a tic-tac-toe board whose squares change depending on the direction you turn a knob, showing a red X, a blue O, or nothing.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

That subtlety matters and says a lot about the project's philosophy. ShiftLens does not turn just any object into an interactive object; it only works where **there is already a displacement to exploit**, or where you accept adding one. This reads less like a flaw than a stance: the interface only exists through the gesture.

## The invisible tool

The least spectacular part of the project may be the most important for its spread: the **design software** that comes with the system.

There is a striking disproportion between the simplicity of the idea and the difficulty of executing it, because printing a lens array at the right curvature, laying out image strips on a backplane and sizing mechanical linkages that move fractions of a millimeter without excessive play each take a skill few people have.

The HCI Engineering group has long experience with this tension between the effect and its tooling. In earlier work such as Photo-Chromeleon (UIST 2019, best paper at the conference), the team had already automated the path from an intention to a reproducible fabrication process, using reprogrammable photochromic dyes to change an object's color.[2](https://hcie.csail.mit.edu/) With Lenticular Objects (UIST 2021), it showed that lenses could be printed on complex curved surfaces to create appearance changes driven by viewpoint.[2](https://hcie.csail.mit.edu/) ShiftLens inherits both building blocks and adds the mechanics: the change is no longer triggered by light or gaze, but by the gesture.

Drawing by hand an array of lenticular lenses aligned with a backplane and mechanical linkages, on a curved surface, without training in optics, is essentially impossible for a non-specialist, so the researchers built a tool that does all that work in place of the user.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The inputs are reduced to a minimum: the images of the desired visual states, the shape and curves of the object. From those few pieces of information, the tool automatically generates the complete ShiftLens structure, ready to be printed in one pass on a multimaterial printer.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The researchers say they thought very carefully about how to communicate the tool's limits to users, a point rarely mentioned in this kind of research. The interface must make clear, without jargon, that ShiftLens does not work with all objects, precisely because it requires a shifting motion between the two layers.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

That choice deserves attention: simplifying the input, automating all the computation, and above all **making the limits visible from the design screen**. Here the tool stops belonging to researchers and starts serving makers, industrial designers or architects.

> From intent to object
> **Four steps, zero electronics**
> - The user supplies the images of the visual states and the object's shape.: 1
> - The tool computes aligned lenses, backplane and mechanical linkages.: 2
> - The object is printed in one pass on a multimaterial printer.: 3
> - The gesture shifts the layers: the displayed state follows the interaction.: 4
> IRZ synthesis from the MIT News ShiftLens announcement.

## Bottle and board

The paper's demonstrators map out the possible uses.

Each demonstrator chooses a different gesture, and that is what makes the series instructive: the bottle exploits a helical rotation that already exists, the screw thread, while the tic-tac-toe board uses a knob whose rotation selects between three states, red X, blue O, or empty square. The system therefore does not force a single interaction grammar; it adapts to the movement the object already has, or to the one you add to it.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The **chemical bottle** best illustrates the value of the approach, since an electronic label on a lab flask would be fragile, expensive and would need power management, while here the bottle only exploits the rotation of its own cap to display the closure state: fully tightened, green check; loosened, red exclamation.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The **tic-tac-toe board** shows that the interaction can have several states and be controlled by a chosen displacement, since the grid squares show an X or an O depending on the direction of a knob's rotation, a small two-position mechanics lesson that requires no electronics to debug.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

Beyond the demonstrators, the MIT announcement mentions more serious applications: **adaptable warning signs** that should survive bad weather, **dynamic packaging** able to alert if fasteners came loose during shipping, and the idea of a **piping network** whose appearance would change to flag a damaged connection causing a leak.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

There is also a visible place for rapid prototyping, for an architect, artist or engineer who wants to quickly explore multi-state objects without going through the sensor-plus-screen-plus-code chain.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

## Stronger than the sensor

Now look coldly at the comparison ShiftLens sets up between a material interface and an electronic one. The paper does not claim mechanics would replace screens everywhere; it shows a specific case where it can do **better than the sensor on its own territory**, that of robustness.

An electronic display is a collection of components, any of which can fail: the sensor, the microcontroller, the battery, the screen, the connections, each adding cost, mass and fragility. The mechanical display of ShiftLens has none of that: no circuit to burn out, no battery to recharge, no sealing to protect components, because the display function **is** the structure of the object.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

That coincidence between function and structure is what gives the system its robustness. Water that would ruin a printed circuit only wets printed plastic; heavy pressure that would crack a screen only deforms an object that was perhaps meant to deform. The displayed state depends on no permanent power supply, since it is carved into the geometry rather than stored in memory.

The approach has another, quieter virtue: it makes the state **readable over time**. An object printed in 2026 will still display its state in twenty years, with no firmware update, no battery running out, no screen going dark, and for laboratory objects, piping or signage, working for decades without electronic maintenance has real value.

## Where it stops

Like any system, ShiftLens has limits that are more honest to name than to work around.

The first is structural: you need a **mechanical shift** to change state, and if the object has no natural movement and you cannot add a button or a knob, ShiftLens does not work. The researchers say so themselves: the system is “not compatible with all objects”.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

The second is tied to the **number of states**, since each state adds more image strips on the backplane and the available surface is finite. Two states are natural; three or four are still possible, but the image becomes narrower and hence less sharp. A screen would shrug off another state, while here every additional color eats display surface.

The third concerns **alignment**. Everything depends on the precision with which the optical layers, the mechanical linkages and the computed geometries coincide, and as soon as parts move too freely or warp during printing, the effect blurs or disappears. The design tool computes the alignment, but the printer must execute it, and shrinkage, warping or tolerance variation from one machine to another remain the classic enemies of multimaterial printing.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

Finally, the system does not guarantee anything about **what is really happening**, since it displays a state derived from a gesture. If the cap is poorly tightened but a slight false contact still turns the ring, the green check appears, so the display works as a mechanical indicator, never a force sensor: it transmits a movement without measuring a strain, and without that distinction the demonstrator bottle gets over-read.

The gap between a lab demonstrator and distributed production adds one more question. The demonstrators were produced on controlled machines, and nobody knows yet whether a consumer multimaterial printer, with its tolerance and temperature variations, would reproduce the required optical alignment. A classic of digital fabrication, and ShiftLens is no exception.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

## Still further

The group's work does not stop there. The researchers announce plans to develop an algorithm able to generate a ShiftLens structure with **fewer user inputs**, and to widen the variety of actuation mechanisms the tool knows how to integrate.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

That second project is more interesting than it looks. Right now, the mechanisms that work are the simply described ones: rotation, sliding, flipping, and opening the palette to more complex mechanisms would make the system usable on objects whose gesture of use is itself complex, like a clamp, a curtain, a hinged lid.

The project sits in a coherent lineage. The team has already published on objects whose appearance changes: Photo-Chromeleon (UIST 2019) reprogrammed an object's colors with photochromic dyes, Lenticular Objects (UIST 2021) printed lenses on curved surfaces for viewpoint-driven changes, and Polagons (CHI 2023) used birefringent materials and polarized light for rotation-triggered changes.[2](https://hcie.csail.mit.edu/) At each step the group chose a new trigger for the appearance change: light, gaze, polarization, and ShiftLens adds the mechanical gesture, arguably the most natural trigger of all, because it is already present in most objects we handle every day.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)[2](https://hcie.csail.mit.edu/) The full paper appears in the UIST 2026 proceedings.[3](https://dl.acm.org/doi/proceedings/10.1145/3746059)

## What it changes for makers

One concrete question remains: what does ShiftLens change for someone who builds objects, a maker, a product designer, an artist?

Today, when a maker wants an object that “reacts”, they are almost automatically pushed toward electronics: a microcontroller, a sensor, a display, code. The path works, but it costs skills and adds fragility, and ShiftLens offers an alternative for a precise class of interactions, those where the state is a simple function of the gesture: tight/loose, open/closed, position A/position B.

The entry barrier is then different, because you do not need to solder, program or design a circuit; you need to know a CAD tool and have access to a multimaterial printer. The ShiftLens tool turns state images into printable geometry, and the result has no firmware to maintain.[1](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)

There is also a durability dimension that speaks to makers. An electronic object becomes e-waste when its battery dies or its circuit burns out, while a ShiftLens object stays an ordinary plastic object, repairable with the same means as any non-interactive object, so the display function does not age faster than the material itself.

## Fabricating the display

At bottom, ShiftLens asks a fabrication question that reaches far beyond the lab: **what if the display were a material rather than a component?**

Every time you add a sensor and a screen to an object, you add a supply chain: a component vendor, a circuit board, a solder joint, a program, a battery, a recycling cycle, and each of those steps is a place where the object can break, stop, become obsolete or cost too much.

ShiftLens offers a different answer, complementary rather than universal: where the gesture of use is already mechanical, the state display can be **fabricated with the object**, in the same printing pass, with the same materials. No extra supply chain, no software to update, no battery to replace.

The green bottle that says “properly closed” is a lab toy, but it demonstrates something concrete: in a world where objects are becoming increasingly talkative through electronics, there is an entire family of interfaces where the material alone is enough to speak. You just have to accept drawing the message in the geometry, and letting the gesture do the rest.

It is also a reminder that 3D printing, too often sold as a machine for making shapes, is first a machine for making **behaviors**. The printed shape is not the finished product; the product is the sequence of states the shape allows, and ShiftLens gives makers one more vocabulary for writing that sequence, using material, movement and light instead of code.

## References

1. [MIT News, “These 3D-printed objects can tell you if they're being used properly”, August 5, 2026](https://news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-you-if-used-properly-0805)
2. [MIT CSAIL HCI Engineering Group, publications](https://hcie.csail.mit.edu/)
3. [ACM UIST '26, proceedings (ShiftLens: Mechanically Actuated Optical Surfaces for Switchable Appearances on 3D Objects)](https://dl.acm.org/doi/proceedings/10.1145/3746059)