No need to worry about the inevitable robot uprising. We will use bacteria to mollify and control robot brains in the future. That's according to a research paper published today in Scientific Reports, which looks at using engineered E.coli to command a robot underling. The experiments -- so far only using modelling techniques -- showed that with a few tweaks, the bacteria could lead to a robot that behaved more like "higher order, more complex animals".
The work was inspired by the impact we already know bacteria has on behaviours in the natural world. This excellent piece from Ed Yong in Discover Magazine describes how the microbiome can have a huge impact on behaviour -- and in one particular case, fruit fly libidos. A study of the insects showed how changes to their gut bacteria impacted sexual preferences. Bacteria in the human gut can also have a huge impact on human brains, professor of anatomy and neuroscience at University College Cork John F Cryan explained earlier this year at WIRED Health.
At the Virginia Tech lab where the study took place, biologists look at using synthetic biology to "reprogramme" the biochemical pathways inside cells -- the same pathways that control decision-making at a cellular level. It's the minutiae of how biological beings operate, with every living cell able to communicate and carry out commands. The team wanted to look at how that decision-making process could not only be engineered, but transplanted to a robotic host using something called a microfluid bioreactor.
The plan is to build small bioreactors where the cells can live, and place them on a robot so that we can have "mobile robots that harbour living colonies of bacteria that direct the robot's behaviour," assistant professor of biological systems engineering Warren Ruder, coauthor on the Scientific Reports paper, said.
But let's take a step back to the slightly less sexy outcome of his study: a mathematical model.
Ruder explains that by using robust equations from bacteria, which we already know accurately describe that bacteria's behaviours, then linking those with equations and models that describe simple robotic behaviours, we can get an understanding of how this kind of setup would work in real life. "When we linked those two we saw emergent behaviour in a virtual simulation -- we were able to watch how a robot driven by bacteria would move in its environment," said Ruder.
The model used the example of a robot that would be fitted with sensors and a tiny microscope, which could take readings from the bacteria on board. When the colour of the bacteria changed between red and green, it told the robot how fast to go, and where, depending on things like colour intensity. This was the first, simple approach, and delivered the expected results. But then the team made a significant tweak. "When we gave the robot one extra component, an ability to talk back to the living bacteria onboard, we saw the cells go from simply toggling between different fuel preferences to approaching a fuel source, pausing and then moving rapidly towards it," said Ruder. This was reminiscent of something Ruder refers to as the "stalk, pause, strike" behaviour that we see in "higher order, more complex animals".
After applying the model to a real world robot-bacteria union, Ruder foresees uses across agriculture (looking at the relationship between soil bacteria and livestock) and oil spills, with drones dropping oil-vacuuming bacteria. If we better understand how bacteria impacts behaviours, as in fruit flies, we could also develop bacteria-based drugs that combat different illnesses.
Even more imminently, Ruder wants to see bacteria-driven robots in schools and universities, teaching the next generation of synthetic biologists about bacteria's massive potential.
This article was originally published by WIRED UK
