Scientists Are Turning Living Cells Into Programmable Machines

Tufts University researchers built Anthrobots, self-assembling machines made from adult human cells with no genetic editing, that swim, repair damaged neurons, and appear to grow biologically younger as they form.

Scientists have built self-assembling machines from adult human cells, and unlike earlier living robots made from frog embryos, these ones get biologically younger as they form. Researchers at Tufts University call them Anthrobots, tiny multicellular structures grown from a single human tracheal cell with no genetic editing involved.

They swim, they repair damaged neurons in lab dishes, and a 2025 study found their cells reverse part of their own aging clock in the process of assembling.

What an Anthrobot Actually Is

An Anthrobot starts as a single adult cell taken from human airway tissue, the kind that normally lines the trachea and helps sweep mucus out of the lungs. Placed in a new environment, that cell divides and self-organizes into a cluster covered in tiny hair like structures called cilia.

Those cilia beat in coordinated patterns, propelling the cluster through liquid the way a tiny paddleboat moves. Sizes range from about 30 to 500 micrometers, roughly the width of a human hair up to the point of a sharpened pencil.

Biologist Michael Levin and then PhD candidate Gizem Gumuskaya led the work at Tufts, publishing the original findings in Advanced Science in November 2023.

How This Differs From the Frog Cell Robots That Came Before

Levin's lab previously built the Xenobots project from frog embryo cells that could move, self-heal, and gather loose cells into replicas of themselves. Anthrobots answer a question that project left open: does this only work with embryonic amphibian cells, or can any cell do it?

The answer turned out to be adult human cells, taken from a patient's own body, with zero genetic modification required. That distinction matters for eventual medical use, since patient-derived material avoids the immune rejection risks that come with foreign or engineered cells.

The Neuron Repair Discovery

In the original 2023 study, researchers scratched a two dimensional layer of live human neurons to simulate an open wound. Anthrobots clustered around the damaged area and formed what the team called a superbot, a fused mass of individual units acting together.

Neurons underneath that superbot showed substantial regrowth compared to untreated wounds. Exactly how the Anthrobots trigger that healing remains unclear, and the researchers were candid that this result surprised even them.

Nothing about the tracheal cells' original biology predicted a capacity for encouraging neural repair once reorganized into a new shape.

Cells That Get Younger When They Reorganize

A 2025 study published in Advanced Science tracked what Gumuskaya and colleagues called the full life cycle of Anthrobots, from formation through natural breakdown. The transcriptomic analysis found something unexpected: assembling into an Anthrobot shifted gene expression toward patterns typically seen in earlier developmental stages.

Measured against standard epigenetic clocks, the cells came out biologically younger after reorganizing than the donor cells they started as. Levin's team is now investigating whether that rejuvenation effect could eventually be harnessed inside the body, separate from any Anthrobot application.

The same analysis showed a shift toward gene expression patterns associated with earlier embryonic development, alongside a measurable self-healing capacity within the Anthrobots themselves. Researchers still do not know exactly what cues trigger that shift, or whether it depends on the specific geometry the cells settle into.

That question connects to a much older puzzle in biology, the one explored in our piece on why cellular aging behaves the way it does at a molecular level.

Why Geometry Instead of Genetics Changes the Calculation

Most synthetic biology works by editing DNA, adding or rewriting genetic instructions so cells behave differently. Techniques like gene editing have transformed medicine this way, but they carry real regulatory weight and long safety review timelines.

Anthrobots take a different route entirely. The same genome that built a person's trachea gets expressed in a radically different physical arrangement, and new behavior emerges from that arrangement alone, with no DNA sequence change involved.

Levin has described this as building with architecture rather than code, closer to rearranging existing hardware than writing new software. That framing sidesteps a lot of the regulatory complexity that comes with direct genetic engineering.

Where This Fits Among Other Living-Machine Research

Anthrobots are one piece of a wider push toward building biological structures without conventional surgery or synthetic scaffolding. Researchers working on 3D printed organs pursue a related goal from the opposite direction, arranging cells layer by layer using a printer instead of letting them self-organize.

Xenobots also went a step further than Anthrobots in one specific way. The frog cell versions demonstrated kinematic self-replication, gathering loose cells into new copies of themselves, a behavior Anthrobots have not been shown to reproduce.

That gap is a useful reminder that these platforms are not interchangeable. Each cell source and design method carries its own capabilities, limits, and open questions that researchers are still mapping one experiment at a time.

What Still Stands Between the Lab and a Clinic

None of this is close to a treatment. Every result so far comes from cell cultures in a lab dish, not from tissue inside a living patient, and the mechanism behind the neuron repair effect is still not understood at a molecular level.

Anthrobots also have a limited lifespan, breaking down naturally within a few weeks under current lab conditions. Researchers still need to determine whether the healing and rejuvenation effects observed in flat neuron cultures would hold up in three dimensional tissue or inside a living body at all.

Anthrobots turn a strange experiment with frog cells into something with a direct line to human medicine, using a patient's own tissue rather than foreign or engineered material. The rejuvenation finding adds a second, unexpected reason to keep watching this field.

Whether either effect, the healing or the reversed epigenetic age, survives the jump from petri dish to patient is still an open question researchers are only beginning to answer.

Frequently Asked Questions

What are Anthrobots made from?

Anthrobots are grown from a single adult human tracheal cell, taken from airway tissue, with no genetic modification involved. They self-assemble into small multicellular clusters covered in motile cilia.

How is an Anthrobot different from a Xenobot?

Xenobots are built from frog embryo cells, while Anthrobots come from adult human cells without any genetic editing. Anthrobots also showed a capacity to promote neuron regrowth that has not been reported in Xenobots.

Can Anthrobots heal damaged tissue?

In lab studies, clusters of Anthrobots encouraged substantial regrowth in damaged neuron cultures. The exact biological mechanism behind this healing effect is still not fully understood.

Do Anthrobots really get younger?

A 2025 study found that cells become biologically younger, based on epigenetic clock measurements, after reorganizing into an Anthrobot. Researchers are still investigating what triggers this shift and whether it could be reproduced elsewhere.

Are Anthrobots used in patients yet?

No, all current research remains limited to lab dish cell cultures. There is no timeline yet for testing Anthrobots in living tissue or human patients.

How long do Anthrobots survive?

Anthrobots typically break down naturally within a few weeks under current laboratory conditions. Their limited lifespan is one of several open questions standing between lab results and any future medical application.