Skip to content

Viruses: Life in the machine

December 13, 2011

An opening figure of my virus chapter shows how biological viruses parallel the behavior of computer viruses. Both kinds of entity infect a host system to propagate their own information, which contains the instructions to spread, infect, and spread again.

Are viruses alive? Biologists have long argued they are not, largely because isolated virus particles are inert, lacking metabolism. But who says life needs metabolism? They use the host cell’s metabolism, as do intracellular bacteria. And giant viruses such as Mimivirus, herpes, and pox viruses probably evolved from cells.

If biological viruses are alive, why not viruses in silico? If life evolved once out of the carbon soup, could it arise again in the machine?

Of course science fiction has long proposed various forms of silicon alien life, such as the Star Trek Horta’s eggs. But an ecosystem evolving out of a machine was not the place they had in mind.

A fascinating approach is to ask: If PC virus “life” has already evolved in the machine, does it show any of the same strategies as those used by living viruses? Let’s look at some ways that computer viruses work.

–PC viruses can be “Fast infectors” versus “Slow infectors.”  Fast infectors infect as many files as possible–like Ebola virus, infecting every kind of tissue. On the other hand, slow infectors only infect files some of the time, or under certain situations. They turn out to be surprisingly effective at propagating. Similarly, herpes viruses are famous for lying low in the body; many kinds rarely affect the carrier at all. Nearly all of us carry undetected herpesviruses.

–PC viruses have ways to avoid detection by the “immune system” of anti-virus software. For example they kill the anti-virus functions (like HIV virus kills human T cells). Or the PC viruses hide within unused parts of a file (like human viruses that persist deep within cells).

–PC viruses can avoid detection by changing their code with each infection. Again, biological viruses do something similar, by allowing rapid mutations in certain genes. HIV virus replication can have an error rate as high as ten percent.

–The best defense against PC viruses is said to be “genetic diversity” of the host population (need I say more).

Some argue that because PC viruses are human-made, they don’t “count” as natural life. However, do we really know this?  Could PC viruses arise spontaneously?

I’m sure some of you know more about the PC angle on this. Are viruses just the beginning; are more advanced forms of life likely to arise in the machine?


  1. December 14, 2011 1:32 am

    It’s very tricky to make computer code evolve because it is extremely fragile. If you change one base pair in a string of DNA, you have probably changed one amino acid in one protein in an otherwise functional organism; it might be a little more efficient, it might be a little less, but it’s probably not instantly lethal. If you change one bit in a string of computer instructions, you might have just changed an instruction from “jump forward 200 instructions” to “add 200 to the value in register C” and broken a major behavior or created a new way to crash the program. So you have to do a lot of work to make code that can evolve without getting selected into nonfunctionality; there are environments that do that, but they’re pretty specialized.

    I expect things like that to arise once engineers start working with evolutionary algorithms to create code by defining constraints and letting the code that solves them evolve in an environment designed to support such development.

  2. shagggz permalink
    December 14, 2011 4:13 am

    I don’t really see computer viruses spontaneously evolving (outside of specialized contexts such as evolutionary algorithms/modeling) due to the error correction methods we have engineered into our information structures, so the error and thereby evolutionary rate will be virtually nil.

    • December 14, 2011 9:24 am

      Surely “virtually nil” can always be quantified? No matter how small the error rate, zooming up the amount of data processed will eventually make errors inevitable.

      Interestingly, biological DNA replication has error correction too. Without error correction, a standard DNA polymerase might have an error rate of one in a thousand; the error correcting enzymes get that down to one in a million. There are many enzymes of DNA repair. When they go wrong, it leads to cancer.

  3. December 14, 2011 9:19 am

    Interestingly, there are DNA mutations where a one point change transforms the entire code. For example, addition or deletion of a base leads to a “frameshift” where all the three-base codons for amino acids are out of register.

    You’re right that most amino acids in a protein are nonessential, so they can be replaced without affecting function of the protein. And most cells have many stretches of nonessential DNA between genes. This is true also of large DNA viruses.

    But RNA viruses like HIV and influenza are different. Their genomes have little nonessential sequence (RNA) and they replicate with many errors. In the case of HIV, maybe one in a thousand virus particles is actually functional at all–that’s the price paid for variability. Influenza virus particles, too, are mostly nonfunctional, because eight segments of RNA have to be packed at random into a particle; and unless you get one of each type, it’s worthless. That is why influenza evolves with huge jumps, where viruses from different host species combine, each contributing a few segments to a scrambled virus.

    So now that engineers work with evolutionary code, do they see dangerous variants arising?

  4. SFreader permalink
    December 14, 2011 11:46 am

    Small differences can indeed make a huge impact ….

    For every one billion particles of antimatter there were one billion and one particles of matter. And when the mutual annihilation was complete, one billionth remained – and that’s our present universe. -Albert Einstein

    Evolving computer viruses are here already, and they can be created using the same rules that appear to govern biological viruses. Major anti-virus manufacturers are on record saying that they anticipate and are working on how to recognize and block evolving malware. (“Evolved” as in not just improved, but evolved as in spontaneously changing.) I wonder whether the shift to cloud-computing will slow down or speed up this evolution.

    Click to access gecco09-sadia.pdf

    The concept of artificial evolution has been applied to numerous real world applications in different domains. In this paper, we use this concept in the domain of virology to evolve computer viruses. We call this domain as “Evolvable Malware”. To this end, we propose an evolutionary framework that consists of three modules: (1) a code analyzer that generates a high-level genotype representation of a virus from its machine code, (2) a genetic algorithm that uses the standard selection, cross-over and mutation operators to evolve viruses, and (3) the code generator converts the genotype of a newly evolved virus to its machinelevel code. In this paper, we validate the notion of evolution in viruses on a well-known virus family, called Bagle.The results of our proof-of-concept study show that we have successfully evolved new viruses–previously unknown and known-variants of Bagle–starting from a random population of individuals. To the best of our knowledge, this is the first empirical work on evolution of computer viruses.

    • December 14, 2011 4:46 pm

      Yes, this report on evolving malware is very cool. And note that antivirus systems fail to detect some of the evolved forms.

  5. December 14, 2011 2:06 pm

    This is one of those awkward little issues. I know a lot of biologists define life in terms of cells and replication. There’s something to this limit, because the other limit is probably that everything is alive in some sense. Stopping between these two extremes gets difficult and arbitrary.

    The issue is that there are many things with most of the properties of cellular life. These range from fire to the biosphere (Gaia–perhaps we’re supposed to be Gaian diaspores?) to songs and memes. Does a blog have a life of its own? How about a political movement? The dreamlines of Australia have lasted through people’s memories for possibly up to 40,000 years. Are they alive or not? Is a superorganism such as an ant colony alive in its own right, or simply an emergent property of a bunch of flightless wasps?

    Oddly enough, the view that everything is alive in some sense is actually among the oldest human beliefs. It’s called animism. And yes, yelling at the screen right now is an animistic gesture. I’m not inside this machine.

    So it’s a bit of a slippery slope. Ultimately, life is a category, and the best definition of life is that it’s what biologists study. Personally I think there’s merit in biologists stepping on the toes of the more egotistic physicists and cosmologists and claiming that life is universal, and that therefore biology is the premier science. However, I have a minority view, I’m afraid.

    On a separate but related note, I’ve noticed an uptick in the US about people getting angry that corporations are defined as persons under US law. My response to this is to quote from the Declaration of Independence: “We hold these truths to be self-evident, that all [people] are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.–That to secure these rights, Governments are instituted among [People], deriving their just powers from the consent of the governed.” Corporations aren’t alive in the cellular sense, they don’t have liberty, being pieces of paper entirely bound by various laws, and they can’t pursue happiness because they do not have free will. Since they cannot possess the inalienable rights of persons, perhaps we shouldn’t define them as persons in law. I’ll be happy when we get a better distinction between the rights of human people and the rights of the legal entities regarded as people for the purposes of the law.

    The point here is that you have to be careful about how you define life. If you get carried away, corporations may become alive. And then what happens?

    • December 14, 2011 4:58 pm

      If we define “life” as something with growth, reproduction, and evolution:

      –Fire is not life, because it doesn’t evolve. No matter how long it goes on, it’s still the same simple redox reactions. And it can arise spontaneously, without a preexisting organism. (Are prions alive?)
      Now, ecosystems have developed that essentially depend on fire as if it were a living player–forests where lodgepole pine trees require high heat to expose their seeds. But fire could also fit the role of inanimate players such as rainstorms and volcanoes.

      –Superorganisms such as ant colonies are alive because the ants are. What you’re getting at is I think a different question: What constitutes an “individual” organism versus a community?

      –Corporations are also a different question, of not just individuality but “personhood.” The “personhood” question is whether you can kill a corporation; does it feel pain? (Maybe; I don’t know much about the legal issues here.)

  6. December 14, 2011 11:38 pm

    I think Tim Allen pointed out that one of the better definitions of life involves having a sharp boundary that designates inside and outside. For organisms, this is cell membranes or cell walls.

    This gets ugly with a superorganism. A single ant (other than a queen) can’t live alone for long, but you can certainly define the boundaries of a superorganism made out of ants. Their cells are just not contiguous masses of cells. Is a superorganism alive or not?

    Fire also gets interesting, because a ball of plasma does have a physical boundary where it differs sharply the air around it. A while ago, I had an interesting conversation with an astrophysicist about whether a dirty, dusty plasma in space could ever become complex enough to be considered alive. The answer appears to be “maybe.” Given the extensive amount of work on building efficient wood-burning stoves right now, it appears that flame is far more complex than we would have thought, given how central it is to human life.

    Corporations are pieces of paper that govern how people contract to relate with each other in a group under a body of laws. Why do we consider them alive again? Oh yes, the law considers them to be persons. That’s the only reason we think they might be alive.

    • December 15, 2011 8:56 am

      The ants are alive, whether superorganism or single. Otherwise, is a human “dead” just because they’re post reproductive age? Furthermore, the isolated cells of ant or human are alive, and reproducing–and can be immortalized in tissue culture.

      The plasma in outer space is interesting–there was Fred Hoyle’s novel, The Black Cloud, which proposed something similar.

  7. SFreader permalink
    December 15, 2011 11:25 am

    In some jurisdictions it is possible to sentence a corporation to ‘death’ (i.e., corporate dissolution) for corporate manslaughter. Something that seems forgotten or overlooked is that most incorporation laws do recognize that the human proxy for the corporation – its Board of Directors – are per law responsible for the actions of the corporation. This means that Directors/individuals can be found criminally culpable, charged, fined and sentenced. It’s largely a matter of will and, increasingly, historical precedent that this is no longer being used to rein in corporate malfeasance. OTOH, civil law requires a lower burden of proof and often awards higher financial penalties than a criminal court — and therefore is likely to ‘hurt’ a corporation where it matters, its P&L. Higher fines despite their never, never collection dates appear to have assuaged the public’s perceived need to also (and equally) reprimand/penalize individual Board members.* Further, some corporations try to limit individual Board members’ responsibility – and personal and financial risk – by agreeing that legal fees related to any CorpX BoD activities will be picked up by the corporation.

    IMO, the biggest problem is that the public has gotten into the habit of thinking corporations/Boards can go unpunished: collectively shrugs and says ‘What I can do about it?’. The laws exist; apply them.

    * The Corporation is the closest human equivalent to an ant colony. Discuss …

  8. December 15, 2011 12:30 pm

    I don’t think a corporation is a superorganism, any more than I think an army, society, or tribe is. While I hate to bring up Godwin’s Law, the Nazis *did* in like the concept of a superorganism for Nazi Germany, so considering human groups as superorganisms is one of those discussions that can get a bit…awkward.

    Actually, I passed up a chance to be on the board of directors of a non-profit group I belong to, simply because I knew I couldn’t afford to personally backstop the group’s finances. I actually think that most corporations (which are small) actually work as designed, with the board of directors being financially responsible.

    The bigger problem with big corporations is the golden parachute phenomenon, where executives get paid whether they screw up or not. To me, this is analogous to parasite load, in that the money and perks soaked up by executives are not going to anything productive. I agree with something Charlie Stross said in Rule 34, about how in a decade, most “rock-ribbed” financial institutions (Wall Street, London, and Shanghai) will be found to be hollow shells, stripped of their assets by their less ethical executives, with the bankrupt remnant left to the shareholders. Some executives may end up in prison, but I suspect that most corporate money will have simply evaporated through “accounting errors” and lost records.

Comments are closed.

%d bloggers like this: