‘We made living robots that self-replicate’

I’ve always been very interested in object making and robots, but who isn’t? Robots are the best. I’ve always thought they were great and wanted to work on them, but it took me a while to get into robotics.


I was very interested in art from the beginning and in high school my focus was more on computer graphics and photography. But when I went to college I started to wonder how good I really was at art and whether it was probably more of a hobby, so I got back into math. After that I worked as a data scientist for an insurance company for three years. It’s not exciting; There are no robots out there!

At 25, I went back to school and studied computer science and robotics. it was what i always wanted to do but it took me a while to realize that i could Do it. It kind of sounds daunting, like you have to be a child prodigy at coding. That’s not true – at least I hope it isn’t!


What is a Robot?

I built my first robot in graduate school, when I was in my late 20s. It was a soft robot made of rubber, which I was interested in during my PhD studies.

What makes a robot different from your laptop computer is that robots move themselves around the world like animals. Like animals, robots can come in all sorts of shapes and sizes, depending on what they are designed to do and the material they are made of.


One of the characteristics of animals is that they are made from soft materials, but we usually make robots out of metal or plastic. Unfortunately, they are not that great because eventually they break down and become garbage and pollution. I started thinking about how we can make robots out of biodegradable and biocompatible materials. Maybe it’s naive, but one solution seems to be to make robots out of the same things animals are made of, which are cells.

My colleagues at the Wyss Institute for Biologically Inspired Engineering at Harvard University and I began to think that soft robots could have many benefits; Maybe they can interact better with humans and stay safe? Or, perhaps they may be more intelligent or perform certain tasks better because they can use softer materials, like animals? But designing soft robots can be difficult so we use AI and computers to help us understand what they should look like.

Usually you have a purpose for a robot: for example, you want it to walk or clean your floor. Then comes the questions of what to make of it, what it will look like and how it is supposed to do the job. “What should my robot look like?” This is a question that most of my work is trying to solve.


Origin of Xenobots

Making robots out of cells actually arose spontaneously from a collaboration. I knew that Xenopus laevis was frogs in our biologist colleague, Doug Blackiston’s lab at Tufts University’s University of Vermont. These eggs are used for all kinds of important developmental biology, and in 2018, Doug bet me we could make robots out of them, so I put in full time.

Skin cells and heart muscle cells were the two simplest to use and the ethics behind it are really good. You take an egg from a day old frog but there is nothing in that egg, there are no neurons and it cannot feel pain. Very often they are thrown away before anything develops. Doug developed all the biology methods for making robots out of cells that are simple enough that I could follow them without previous biology training, and then we asked how we could make this robot do what we want. want to do.

That’s where AI comes in, it’s more of a collaborative and design tool. It offers all these different ways of putting these cells together to make a robot that you want; Walking, for example. It’s a lot like using Lego blocks, but there are Lego cells and tissues.


The computer puts cells together in random ways and determines which of these configurations does more or less what we want from the robot. As you can imagine, a random group of cells is probably not going to do what you want. Some won’t move at all, but some may fall down, which is a bit close to walking or running along. The computer takes those good designs, modifies them and removes the bad designs.

Once we showed that we could design living robots—which we called xenobots—to behave in specific ways, we were then able to see that they could do all these amazing things that regular robots can do. It’s not that easy to do.

Finding out why xenobots self-reproduce and self-heal

Xenobots can heal themselves, which is surprising from a robotics point of view – although perhaps not so surprising to a biologist, since we all heal ourselves. You can almost cut this robot in half and it just zippers up and starts moving like before. No other robot can do this. Some of these behaviors that seem obvious to biologists turn out to be truly amazing ideas and perspectives to roboticists.

But the way xenobots reproduce was surprising to both roboticists and biologists. Every single organism on the planet that we know of, reproduces either sexually or asexually. Reproduction involves a parent or parent organism that gives part of itself, and from that small part, or parts, a child or offspring grows.

We saw that there was something really amazing that Xenobots could do. If we sprinkle loose stem cells into their dish the xenobots will spawn in the dish, pushing those cells together into piles that develop into more xenobots. They are able to reproduce in a completely different way—it was like the xenobots that make up other xenobots.

When we gave their children more cells to make, they sometimes produced grandchildren. But then breeding stopped. So, we wanted to see if it could last longer.

We asked our computers what we could do to increase reproduction and make it last longer, and the most successful change was changing the shape of the original Xenobots. You can carve them into any shape you want, even a small dog! But after weeks of trying different sizes, it gave us a very simple design that looks like Pac-Man. This xenobot has a carved “mouth” and moves around, and because of this, it traps cells there and is better at being a snowplow; Pushing cells into large piles. These larger piles are more likely to develop into offspring and this process of self-replication lasts longer.

Living Robots That Can Self-Replicate
Forty-nine AI-generated parent Xenobot shapes that repeat more often than the default circular shape.
Douglas Blackiston and Sam Craigman

In 2020, we discovered that we could make robots out of cells and began to see the earliest evidence of this self-replication. But it took some time to make sure it wasn’t just chance and to prove that Xenobots was indeed building other Xenobots, so we didn’t release our findings until late 2021.

It’s a lot easier than most people realize, and it’s probably important to emphasize that. A xenobot is not the most complex and useful robot, there is a lot it cannot do and there are a lot of conditions that have to be perfected. However, roboticists all believe that it is a robot and a kind of self-reproduction. This isn’t a robot dancing on the internet or building out, it’s pretty limited, but our description is accurate.

Will the surviving robots take over the world?

People ask are Xenobots going to take over the world? We are very confident that this technology is safe because they cannot reproduce unless they are in a petri dish and we keep giving more cells to make them. Even if we do this, the self-replication stops after about five rounds. And that too took months; It was a lot of effort and still humans are in the loop. It is also extremely safe as it is just frog skin cells. Frogs shed their skin in lakes and streams all the time and we really don’t care. This is within a lab and there are very strict policies in place. But, I think it’s healthy for people to be skeptical about this kind of science and technologies getting out of control. I also understand that there is something in this that might upset people; This is a robot made of frog skin. But I think if you dig a little deeper, you can see how easy it is and how well it can do it.

We’ve been building robots made entirely out of only living cells for three years, but in many ways they are already much better than the robots we’ve been building since the 1940s. These are self-healing, biodegrading and self-replicating robots.

Right now, xenobots will have to live in an aquatic environment. So, if they’re going to be useful robots, they’ll probably have underwater applications. They may clean up lakes, streams, and oceans, because waterways have many tiny contaminants that are difficult to see. We know xenobots are great at building stacks, and they are self-powered as well. They are pre-loaded with fat and protein which they burn. They live off that energy for about two weeks and then they biodegrade.

Sam Craigman develops self-replicating robot
Living Robots That Can Self-Replicate

Science is not easy, most experiments do not go as expected. You fail 100 times in a row, then the 101st experiment is really cool and the world can take notice, which happens sometimes and it’s amazing. It is really beneficial. It’s overwhelming at first, but of course I’m very grateful for all this support and interest. A lot of science research goes unrecognized, which is sad.

I hope this is exciting to people. In particular, we have seen youngsters getting extremely excited about Xenobots. I get dozens of emails every month from students from middle school to undergraduate, and I’m passionate about getting the next generation involved in this work. We are making soft robot kits for the classroom and recently we are trying to do the same with Xenobots. We hope to make a really simple kit so that many people can see for the first time that it is nothing to worry about since they made it in their biology class.

You can learn a lot from it, and who knows what the next generation will do with it. Maybe more than I can imagine now.

Sam Kriegman PhD is a computer scientist with a joint postdoctoral appointment at the Vice Institute at Harvard and the Allen Discovery Center at Tufts. His work on Xenobots is developed with senior scientist Doug Blackiston Ph.D. You can follow Sam on Twitter @crigmerica,

All views expressed in this article are those of the author.

As told to Jenny Howard.