Level 1 — Absolute Beginner

Scientists at MIT in America built a very small robot. The robot is flat and thin. It is about as long and wide as a stick of chewing gum. It is only half a millimetre thick.

This robot is special. It does not use a motor. It uses real muscle cells from an animal. The cells are alive and they sit on top of a soft gel.

The scientists changed the cells. Now the cells move when light touches them. The robot has two fins. When light shines on one fin, that fin flaps. The robot swims forward in water.

The team moved a light around a dish of water. The robot followed the light. It swam through a simple maze. It is slow. It swims four body lengths in one minute. A good human swimmer is much faster.

robot

A machine that can move and do tasks.

flat

Thin and level, not thick or round.

muscle

The soft tissue in a body that makes movement possible.

cell

The smallest living unit that living things are made of.

gel

A soft, wet material that is thicker than water.

fin

A flat part that a fish uses to swim.

flap

To move up and down quickly.

maze

A set of paths that is hard to find a way through.

Level 2 — Elementary

A team at MIT has made the first flat robot that can swim using living muscle. The work was published on September 29, 2026 in the journal Advanced Functional Materials, under the title 2D skeletal muscle thin film actuators enhance efficiency of biohybrid robots.

The body of the robot is a soft gel film called GelMA. It is about the length and width of a stick of gum and roughly half a millimetre thick. The two halves of the film act as two fins. Each fin is covered with one single layer of live skeletal muscle cells, a layer thinner than one human hair.

The cells have been genetically engineered so that they contract when light shines on them. If the team lights one fin, that fin flaps and pushes the robot through the water. If they alternate the light between the two fins at chosen intervals, they can control both direction and speed. Before testing, they even exercised the muscles by flashing light at them, like a training routine.

At full speed the robot travels about four body lengths every minute. That sounds slow next to an Olympic swimmer, who can cover up to 65 body lengths a minute, but it is roughly the pace of a cow shark. Ritu Raman, the associate professor of mechanical engineering who leads the lab, explained the difficulty simply: it takes a lot of force to move through water versus air. The team hopes machines like this could one day monitor rivers and oceans, or handle delicate tasks that normal hardware cannot.

journal

A publication where scientists report new research.

film

A very thin layer or sheet of material.

skeletal muscle

The type of muscle attached to bones that moves the body.

genetically engineered

Changed on purpose by altering an organism's genes.

contract

To become shorter and tighter, as a muscle does.

alternate

To switch back and forth between two things.

interval

A period of time between two events.

monitor

To watch or measure something over time.

Level 3 — Intermediate

Most soft robots that borrow their power from living tissue have been three dimensional: a block or tube of muscle wrapped around a scaffold. A team in Ritu Raman's lab at MIT has now demonstrated the first two dimensional version capable of locomotion, and the result swims. The study appeared on September 29, 2026 in Advanced Functional Materials as 2D skeletal muscle thin film actuators enhance efficiency of biohybrid robots, with graduate student Maheera Bawa as first author and Arielle Berman, Laura Schwendeman, Ferdows Afghah and Seanbiron Johnson as co authors. The Office of Naval Research provided part of the funding.

The design is deliberately minimal. A film of GelMA hydrogel, about the footprint of a stick of chewing gum and some half a millimetre thick, serves as the skeleton; its two halves function as fins. Over each fin lies a single monolayer of skeletal muscle cells, thinner than a strand of hair, modified with light sensitive channels so that they twitch on illumination rather than on electrical stimulation. Shine light on one fin and it flaps; alternate the illumination between fins at controlled intervals and the direction and velocity of the swimmer can be dictated from outside the water entirely.

Getting there required unglamorous materials work. The team replaced the fibrin used in earlier generations with GelMA, then tuned composition, stiffness and the geometry of the microgrooves that guide cell alignment. Square bottomed grooves aligned the cells better than curved ones, and stiffer gels generated more force. They also trained the tissue, flashing light at the muscle on a schedule that functions as an exercise regime before performance testing.

The performance is modest and honestly reported. Top speed is about four body lengths per minute, against as many as 65 for an Olympic swimmer, though that puts the robot roughly level with a cow shark. In a large petri dish the researchers moved a light source by hand and the device followed it, threading a simple aquatic maze. Raman framed the engineering obstacle plainly, noting that it takes a lot of force to move through water versus air, and identified the next milestone: optimising the body design to enable faster swimming. The longer ambition is machines soft and biocompatible enough to monitor fragile aquatic environments, or eventually to perform delicate jobs in settings too unpredictable for conventional hardware.

scaffold

A supporting structure on which cells or tissue are grown.

locomotion

The ability to move from one place to another.

hydrogel

A water rich jelly like material used widely in bioengineering.

monolayer

A single layer of cells or molecules, one unit thick.

illumination

The act of shining light on something.

velocity

Speed in a particular direction.

Level 4 — Advanced

Biohybrid robotics has spent a decade building actuators out of living tissue, and almost all of them have been volumetric: muscle grown around or within a three dimensional scaffold, with the attendant problems of nutrient diffusion and fabrication complexity. The contribution from Ritu Raman's laboratory at MIT, published on September 29, 2026 in Advanced Functional Materials under the title 2D skeletal muscle thin film actuators enhance efficiency of biohybrid robots, is to collapse the architecture into a plane and show that the plane can still swim. Maheera Bawa, a graduate student in the lab, is first author, with Arielle Berman, Laura Schwendeman, Ferdows Afghah and Seanbiron Johnson as co authors and partial support from the Office of Naval Research.

The device is almost provocatively austere. A GelMA hydrogel film, approximately the footprint of a stick of chewing gum and on the order of half a millimetre thick, constitutes the entire skeleton, its two lobes serving as fins. Each lobe carries a single monolayer of optogenetically modified skeletal muscle cells, a sheet thinner than a human hair, which contracts on illumination rather than on electrical pacing. Because actuation is optical, control is fully untethered: directing light at one fin produces unilateral flapping and therefore turning, while alternating illumination between fins at prescribed intervals sets heading and velocity without any physical connection to the swimmer.

What enabled the reduction in dimensionality was materials optimisation rather than any conceptual leap. Substituting GelMA for the fibrin of earlier generations, the group swept composition, stiffness and microgroove geometry, finding that square profiled grooves produced markedly better myotube alignment than curved profiles and that stiffer substrates yielded greater contractile force, a trade off against the compliance usually prized in soft robotics. The tissue was additionally conditioned, with scheduled optical stimulation functioning as a training regimen that improved output before characterisation. Reported peak locomotion is roughly four body lengths per minute. Raman's own benchmark for that figure is unsparing, noting that an Olympic swimmer manages up to 65, and that the honest biological comparator is the cow shark.

The demonstration experiment was correspondingly simple: a large petri dish, a manually translated light source, and a device that tracked it through a rudimentary maze. The significance lies not in the speed but in the efficiency claim embedded in the title, since a planar actuator avoids the diffusion limits that constrain thick muscle constructs and is far easier to fabricate reproducibly. Raman's stated framing of the core difficulty, that it takes a lot of force to move through water versus air, identifies exactly where the design must improve, and she has named body geometry as the next variable. The applications she sketches are consistent with the hardware's properties rather than inflated beyond them: environmental monitoring in aquatic settings, and in the longer term delicate interventions in environments too fragile or unpredictable for conventional machines, with commentators also raising precision microsurgery as a plausible destination.