How the xenobots were made is more important than the robots

You might have seen this news - Scientists use stem cells from frogs to build first living robots or Meet the xenobot: world's first living, self-healing robots created from frog stem cells. Both report the creation of biological robots (biobots) built from cells scraped from the embryos of African clawed frogs, Xenopus laevis (hence the name xenobots).

It is a big step forward for biorobotics - the branch of robotics that builds living machines. But I want to bring your attention to the method used to create the xenobots.

Those robots were designed in silico - inside a computer. Researchers leveraged AI and evolutionary algorithms to find the right design for xenobots.

In a simulated environment, researchers tested many different designs against each other. The algorithm then picked the best performing designs and created a new iteration of xenobots which then were tested again in the simulated environment and then the cycle started again. This method resulted in a design that did what researchers wanted them to do with minimal manual input.

Then the robots were made in real-world and voila! we have xenobots.

This process can be highly automated. That's exciting. I can imagine a system where you specify what you want you biobots to do and then the computer runs thousands of simulations for you. From time to time you check if the design goes in the right direction and adjust the simulations (I can see here reward modelling being applied). Knowing AI and evolutionary algorithms, you might get something that you completely did not expect but it does the job.

When you are happy with the robot, the 3D bioprinters go to work and a couple of days later you get your custom biorobots delivered.

The basic idea and building block are here and now it is just a matter of putting all the pieces together and getting rid of rough edges.

I'm excited to see biorobotics growing. This field has huge potential. It can create microscopic machines to clean the environment, deliver drugs, sort trash or be parts of a microfactory. I can't wait to see the next innovation to come out of biorobotic labs.

The paper: A scalable pipeline for designing reconfigurable organisms