Strings of binary aren’t all that different from strands of organic DNA; they both carry actionable information encoded into reconfigurable symbols. And, like DNA, with enough replication and slight variations, software could become resistant to viral attacks through digital biodiversity.
Taking inspiration from nature, scientists at the University of Denmark’s Center for Fundamental Living Technology (FLinT) devised a method that allowed information strings made of binary code to autonomously self-replicate and mutate in a virtual simulation. Basically, they got digital strings of 1s and 0s to act like the building blocks of organic life.
According to the researchers, this finding constitutes a step toward understanding how digitized information—knowledge and software—can ensure its own survival over time by continually generating variable copies of itself, like our DNA does, preserving valuable data indefinitely. As long as it has a physical container capable of computation, anyway.
“In the real world, everything falls apart. The mountains fall apart. Things deteriorate. For such a system to work, you need to be able to maintain long polymers—long molecules—that contain information,” Steen Rasmussen, the head of the FLinT center, told me. “I think that these autocatalytic, or self-maintaining, networks are very robust. So you can perturb them, you can take away some of the components, and they will immediately be regenerated.”
The team’s approach, described in a paper published in Europhysics Letters, involved creating a virtual pool of information strings (combinations of binary numbers, or “polymers”) made to act like the ingredients of a chemical reaction.
In the simulation, a binary polymer replicated itself by joining two matching bits of binary, mimicking DNA ligation—when two DNA fragments team up to make a larger molecule—and mutated itself with another random joinder. Mutations that degraded the information string were filtered out.
The result: long strings of code organized into reliable patterns, with the ability to preserve themselves theoretically forever. According to Rasmussen, the number and concentration of whole information strings his team’s approach yielded is novel, indicating its promise for engineering digital biodiversity.
“With this particular set-up, the system makes it so that all polymer strings have a similar concentration—there’s just as many long molecules as short molecules,” Rasmussen explained. “I’ve been working in this area for many years, and I was very surprised when I saw that.”
The work is certainly technical and quite narrow, but it’s a branch of research that fits into FLinT’s wider goal of creating living machines: devices that can repair themselves, and even replace themselves. A base example might be a 3D printer that can also print a copy of itself, or a piece of code that can proliferate and preserve itself like strands of DNA.
“There’s no doubt in my mind that living and intelligent technology will be the next really big thing the same way that we’ve had a revolution with information technology,” Rasmussen said. “We’ve only seen the tip of the iceberg.”
Computer science research that uses nature’s biodiversity as a guiding principle is a burgeoning area. Security experts are looking at ways to make every piece of software unique by engineering slight variations in their code, though via decidedly less complex means than Rasmussen and his colleagues. George Church, an experimental geneticist, has spoken about a future where data is coded into plant DNA and grown in the wild, giving data the survival edge of biodiversity in more literal terms. To date, he’s encoded his 2012 book, Regenesis, in DNA.
The idea of digital code that reproduces like organic DNA conjures up all sorts of science fiction imaginings and philosophical questions about the nature of humanity and technology. FLinT’s own site makes no bones about the possible applications of their research as having “the character of science fiction and not science fact.” It’s important to note that the team’s findings are preliminary and experimental, and much more research is needed before it can be applied.
“This is not a result that I can say, okay, let’s go turn the cranks in the lab and do something specific. This is a finding that says, wow, it is indeed possible to make long polymers that are very highly selective. We didn’t know that before,” Rasmussen said. “This is a brick, and there are many bricks that go into a house.”