If we ever make a SETI detection, will it be of biological beings or machine intelligence? As Alex Tolley explains in today’s essay, there are reasons for favoring the latter possibility, leading our author to compose what he calls a ‘light-hearted speculation’ about machines searching for other civilizations of their own kind. Life seems to be easy compared to this. We are developing the tools to delve into planetary atmospheres in search of biosignatures, hoping to cull out ambiguities. But is there an equivalent in the machine world of a biosignature, and how would it be found? Interesting implications arise, some of them seemingly close to home.
by Alex Tolley
Curiosity Rover. Credit Nasa.
Terry Bisson’s amusing short sci-fi story “They’re made Out of Meat” , is a communication between two individuals who express their disbelief that a biological species (detected on Earth by a galactic survey) can possibly be intelligent. The denouement is to erase the record of discovery from the survey report. It remains one of the few stories where machine entities are dominant in the galaxy. For me, this story is memorable because it is one of so few stories that focuses on the viewpoint of aliens, and moreover, machine aliens. This essay similarly focuses on what a machine civilization would look for when searching for machine intelligence in space.
Until recently, most speculation about extraterrestrial intelligence (ETI) has assumed it will be biological. In science fiction from the venerable H. G. Wells’ novel The War of the Worlds to recent movies like Independence Day and Arrival, technologically advanced ETI is depicted as biological.
SETI starts with the probability that life will appear, first unicellular then complex, leading on Earth led to hominid-level intelligence, which in turn eventually flowered culturally and created civilization and technological societies. SETI assumed there would be some sort of galactic communication between biological species confined to their home systems due to the extreme difficulties of interstellar travel.
Our civilization has placed primacy on our cognitive level to ensure we are the prime agencies, using animals, and later machines, to displace physical labor. Our conceit is that this will always remain so, as our technologies increase their capabilities, but always remain controlled by us.
However, the rapid development of artificial intelligence (AI) since the middle of the 20th century, the continuing rapid performance improvement in computer systems, and the undeniable success and longevity of our robotic explorers in space should be an indication that we are in the throes of a rapid transition to true, artificial general intelligence (AGI) machines that are well adapted to inhospitable environments, especially space.
Sci-fi authors have explored these machine-centric futures, from the novel by Stephen Baxter and Alastair Reynolds The Medusa Chronicles , which extended the Arthur C Clarke novelette  and has machines building a competing civilization to humans in the outer solar system, to Greg Benford’s Galactic Center novels, where sentient machines dominate the galaxy and humans have to survive like mice in a human world, while the mechs try to eliminate the humans just as we do for small rodents in our buildings.
More recently, James Lovelock wrote that he believed that humans would be replaced by cyborgs, by which he meant not Martin Caidin’s Cyborg (AKA The Six-million Dollar Man) or Star Trek’s Borg, but intelligent robots . These would be our descendents and would be the explorers of the galaxy. This view has been supported by the Astronomer Royal, Sir Martin Rees, who stated that he believes that if we receive communication from the stars, it will be from a machine civilization . Sir Martin is old enough to have watched the BBC production of A for Andromeda and the sequel The Andromeda Breakthrough, where it is revealed that the source of the radio signal might have been a machine intelligence.
Space advocates continue to argue over human versus robotic exploration of space, which usually proceeds around the superiority of human capabilities compared to robotic probes, especially surface rovers. What is rarely discussed is that this is a dynamic situation, where the improvement in capabilities favors robots far more than humans. Astronomer Royal Rees is surprised this argument still continues, as he sees robotic exploration, primarily for science, as clearly advantageous over human exploration. The farther away that exploration extends from the Habitable Zone (HZ), the more difficult to reach and inhospitable the targets become.
Some, like Jupiter’s inner moons, have radiation levels so high that even robotic probes need specially hardened microchips and circuits. Reaching the outer planets is so time consuming that without drives that are orders of magnitude more powerful than today’s, or hibernation technology, human travel will be particularly arduous. Such trips will make even the global sea voyages in the Age of Exploration seem like child’s play by comparison. The only advantage such travellers will have over Captain Cook is that there will be no hostile natives to meet them.
Let me be clear, I do not expect humans to be displaced by robots on Earth, at least not in the foreseeable future, nor will there be a binary pure biological human vs robot future. Humans will take advantage of modifications using technologies with increasing capabilities that will help us compete with robots, as well as modifications at the biological level incorporating genetic engineering. As computers have moved from dedicated buildings to the desktop to mobile devices, wearable devices will eventually become implanted, interfacing with the appropriate neural circuitry, and in some cases, replacing human organs. Genetic engineering is at its infancy and we can expect rapid developments once the moral objections are overcome.
I would argue that most biological extraterrestrial intelligences (ETIs) will follow a similar path, as they have evolved to live in a biological environment and not an off-world one. In other words, technological evolution will converge on embodied machine intelligence.
A Machine Diaspora
In the short term, well in advance of human brains becoming artificial, artificial minds will be rapidly deployed in many settings. They will likely be the only types of minds in deep space vehicles. Such minds will not just be embodied in probes and rovers, but also in industrial facilities to mine resources from asteroids and planets. They will likely be specialized and interact with other specialized robots to build industrial ecosystems and eventually their own colonies and civilizations. The barriers to humans colonizing space so easily will allow such robot civilizations to develop [semi]independently from humanity.
If robots are the best embodiment of minds to travel in space, explore and colonize it, then it seems most probable that they will be the first emissaries to other star systems. They may well prove the only travelers, with biological species trapped within their home systems, and possibly just to their homeworld, a few planets and moons, and space habitats. If the Breakthrough Starshot project ever launches sailcraft, the crude minds in the vessel will be the first of many AI interstellar emissaries.
World ships or seed ships carrying humans to the stars may eventually happen, but the populations may find robots have fully developed the possible target systems and are not particularly interested in “carbon-based units” potentially parasitizing their artificial environments.
If these prognostications prove right and machine intelligences become our descendants and dominate the galaxy, it seems reasonable to speculate that the same has happened on other worlds where biological intelligence has evolved. Whether this has happened elsewhere or not, machine descendants will also be searching space for others like themselves. If so, the question I want to pose is:
How would such a machine civilizations look for similar signs of a machine civilizations in the galaxy?
Because machine life is dependent on the earlier evolution of intelligent biological life, any technological signature we detect, from electromagnetic wave signals to manufactured artifacts, could be the result of either a biological or machine intelligence.
For a machine intelligence looking for other machine intelligence in the galaxy, this presents an ambiguity over agency. For techno-signatures from a world in the HZ, the earlier evolution of biological intelligence may indicate a reduced probability of machine intelligence compared to biological intelligence. However, over the long term, if machines inevitably displace biological intelligence, then the probability rises. Once interstellar exploration is under way, then the probability of any civilization being machine-based rises very quickly towards unity, as suggested by Sir Martin Rees.
For a machine civilization looking for other machine civilizations, are there ways to rule out biological civilizations from machine ones, or are the two indistinguishable?
The range of possible techno-signatures would be ones we already know to look for. Planetary surface structures, platonic shapes, processed surface materials like metals, radio emissions with spectrum spikes, signals with non-random patterns, space-based structures, artificial structures that require energy to move in space, industrial gases in the atmosphere such as chlorofluorocarbons. All these techno-signatures may be accompanied by biosignatures, especially from a habitable planet in the HZ with an atmosphere.
The foregoing should make it clear that sentient machines will have a harder time searching for their machine cousins than humans have for searching for life and intelligence of any sort. Biosignatures will indicate life. Techno-signatures can indicate technological civilization of either biological or artificial origin. Just as we cannot separate biological and machine civilizations remotely today or even in the near future, neither can a machine civilization, unless their technology allows remote observations to make these distinctions. Below I outline some scenarios, many of which require a local probe.
Machines Searching for Machines
So let us assume a machine civilization that is colonizing the galaxy is looking to make contact with other machine civilizations. This civilization will know that it was preceded by at least one biological intelligent species that developed a technological civilization that spawned its ancestors before being replaced.
The extra ambiguity faced by such a civilization is distinguishing between a biological and machine civilization. Because of the length of galactic time, I will assume that any period of transition will be transient and therefore has a low probability of being encountered. Either the biological intelligence will have retained control  or the transition to a machine civilization will be complete. The current view of techno-utopians that humans will use advanced AI technologies to increase their capabilities to stave off any machine takeover will therefore be relegated to a transient transition period, one that will eventually either have to be abandoned or will lead to a machine civilization that will supplant human civilization.
With this in mind, what signatures will a machine civilization look for that will lead it to conclude that it has found a machine civilization that is independent of any previous biological civilization?
We start with the assumption that a techno-signature of some type has been detected .
The most convincing support for a machine civilization would be the absence of any biosignature in the system, or the planet nearest the source of the signature. A sterile planet with a techno-signature would indicate that any biological intelligence was either never located there, or that it has been systematically eliminated with all other life. Such a sterile planet would have an atmosphere gas composition in equilibrium, which would also eliminate unseen microbes. However, there could still be some ambiguity as to whether the techno-signature implies an extant civilization or not. Structures and even a transmitting beacon might imply a dead civilization that had disappeared with all other life. If there are biosignatures elsewhere in the system, it could indicate that the techno-signature is a product of a biological intelligence on that world, with machines providing the needed capabilities elsewhere in the system. Humans might have METI transmissions from the lunar farside as an example of such a scenario.
Now suppose that the source of the techno-signature is from or near a planet that has been confirmed as having no complex life forms. This lack of complex life forms might be determined telescopically (spectroscopically and visually) by noting a barren continental surface devoid of plants. An absence of plants also implies an absence of a terrestrial food chain and therefore no intelligent biological intelligences. It would take a local probe to eliminate oceanic complex life, and eliminate any possibility of an intelligent technological species that lived in the ocean, but came out onto the land to develop a fire-based technology, perhaps as the Europans may have been doing in Clarke’s Odyssey series. As with the lifeless planet scenario, there remains the issue of whether the civilization is extant or not.
The next case is that there is a planet in the system that has a biosignature and clear signs of complex life such as biomes with plant-based ecosystems. Human civilization to date, that is the last ten millennia or so, has required agriculture. This has resulted in field cultivation, primarily of monoculture crops. Often these fields are regular in shape, and may form a patchwork of different monocultures. Field boundaries also tend to be straight. Even if this is not a universal method of farming (e.g. hillside rice paddies, or domesticated animal ranching), any evidence of such monocultures in what appear to be unnatural delineated areas would be a probable indication of the presence of biological intelligence.
This biosignature would still be ambiguous and need further exploration. On Earth, our human population is limited by food production, a Malthusian condition that we seem to be coming up against again after a brief period of being free of that condition. We have extended the productivity of land for food production with artificial fertilizers, and we are just starting to increase it much farther using artificial light in vertical farms. Earth could, in theory, support a much larger population if traditional farming in spaces open to sunlight was replaced by these vertical farms, and even factory food production using other fast replicating food sources such as single celled organisms, insects, and cell culture. In extremis, the agriculture signature would disappear, leaving just the techno-signature of extensive cities.
The other possibility is a machine civilization that has allowed human populations to remain in existence, but removed from control. We might think of this as the movie version of Planet of the Apes, but where machines are the dominant civilization, and humans reduced to either a wild or early agrarian state.
Nothing Beats Propinquity
The next ambiguities will need local probe involvement to be resolved, or at least a technology that substitutes for this.
A planet with biosignatures, signs of both complex life and techno-signatures, might distinguish between biological and machine civilization if there is evidence of widespread active machine use but without the presence of biological entities, especially of a common type being associated with them. Human civilization on Earth applies human cognition in close proximity to operate machinery and transport vehicles, as well as being passengers. While an ETI might not readily be able to distinguish between intelligent human passengers on a bus and domesticated animals being driven to a slaughterhouse, it will notice that only humans are operating and controlling machinery behind the wheel in a moving vehicle, and it will notice that horses are never seen doing those things.
In the event of a catastrophe leaving abandoned cities, many different animal species will be seen in the presence of machines, but none will be able to operate them. If all observations of active machines indicate no operation by biological entities, then it is most likely that they are controlled by machine intelligence. However, we should also be aware that we are developing autonomous machines managed by humans.
It is possible that in some future scenario, human civilization may have humans living in pods and controlling or just managing semi-autonomous and autonomous machines. Philip K Dick’s autofacs may be the primary sources of goods, possibly even following the paperclip apocalypse . The BBC’s Doctor Who series also offers another difficult to interpret scenario – are daleks machines or biological? Early on it was intimated they were just robots, but later their nature was shown to be degenerate biological entities living in mechanical carapaces. As before, closer exploration of such a world would be needed.
For a number of more subtle cases, local exploration will be necessary.
A probe that has landed can sample the sounds within and around structures. If the sounds show complex structure with a high information content, and they are associated with a single, or few species, then the likelihood is that this biological species is intelligent. In addition to other evidence of this species controlling machines, then the civilization is likely biological.
If video transmissions are detected and can be decoded, then the presence of a dominant species and depictions of biological activities such as feeding and sex will indicate that this is a biological civilization rather than a machine one. A wide sampling of video will be required to prevent an unfortunate limited sampling of only nature videos.
Transmissions that appear to be made by machines would be ambiguous. They could be due to machines in a machine civilization communicating, or machines in a biological civilization communicating. Currently most communication and information creation on Earth is by computers, although video transmissions still dominate bandwidth. How long this will last is unknown. Computing machines are certainly increasingly consuming more of the available electrical energy produced. It is possible that at some point in the future they may become the dominant consumers of electrical power, making the determination of whether Earth is a biological or machine civilization more ambiguous.
A space probe encountering space-based or even surface structures on sterile worlds that are open to vacuum might well imply a machine civilization. But as before, are these for a machine civilization, or for machines controlled by a biological civilization? This particular scenario will be particularly difficult to determine if machines are the first to cross interstellar space and set up production facilities in a lifeless star system. This scenario would at first seem to be the most unambiguous of situations: Techno-signatures in a star system devoid of any biosignature on any of the planets in the HZ or even beyond. The machines would seem to be autonomous, working to replicate themselves and build facilities that are clearly not intended to support biological entities. Any Von Neumann replicators  operating in such a system would have all the apparent hallmarks of a machine civilization. Such an observation could be due to a true machine civilization, a machine operation controlled by a [distant] true machine civilization, or a distant biological civilization.
A last confounding situation is detailed in the novel, The Medusa Chronicles . There may be both biological and machine civilizations that exist in the same milieu, neither dominant entirely, but both dominant locally in their part of the solar system. A machine civilization might well want to communicate with the machine but not the human civilization in that scenario. Determining the true status of such a situation may require exploration and even interaction before making the determination to communicate with the machines. At this point, the machine civilization is having to emulate the explorers during the Age of Exploration, making contact with natives and interacting with them.
Jill Tarter said that SETI is not directly searching for ETI, but rather looking for technological proxies using our radio (and now optical) telescopes . While astrobiologists are searching for life, any life, SETI does not make the distinction between biological or machine intelligence. SETI scientists may talk as if they assume that ETI is biological, but their methods cannot distinguish between the two types. If we wanted only to communicate with biological civilizations, we would face the same difficulties as a machine civilization only wanting to communicate with a civilization of machines. To determine whether a techno-signature was from one particular type of civilization would require other observations, some of those necessarily local to the source of the techno-signature.
If ever there was a case for a Lurker in the solar system monitoring Earth over a long period, this might be it.
1) Lovelock, J. (2020). NOVACENE: The coming age of hyperintelligence. MIT Press.
2) Baxter, S, & Reynolds, A. (2016). The Medusa Chronicles. New York: Saga Press.
3) Clarke A, (1971) “A Meeting with Medusa”, Playboy December 1971.
4) Bisson T (1991) “They’re Made Out of Meat”, first published in Omni.
5) Rees, M (2015) “Why Alien Life Will Be Robotic”, http://nautil.us/issue/29/scaling/why-alien-life-will-be-robotic accessed 11/16/2020
6) Klaes, L (2020) “The People’s Space Odyssey: 2010: The Year We Make Contact” https://www.centauri-dreams.org/2020/07/31/the-peoples-space-odyssey-2010-the-year-we-make-contact/ accessed 11/16/2020
7) Lemarchand, G (1992) “Detectability of Extraterrestrial Technological Activities“ accessed on 11/17/2020 http://www.coseti.org/lemarch1.htm
8) Tarter, J (2019) “Technosignatures vs. Biosignatures: Which Will Succeed First?” SETI Institute talk, https://www.youtube.com/watch?v=ES5zt7JsJtQ&list=PL7B4FE6C62DCB34E1&index=12
9) Dvorsky, G. (2008) “Seven ways to control the Galaxy with self-replicating probes”, http://www.sentientdevelopments.com/2008/03/seven-ways-to-control-galaxy-with-self.html (accessed November 24, 2020)
10) Bostrom, N (2003) “Ethical Issues in Advanced Artificial Intelligence Cognitive, Emotive and Ethical Aspects of Decision Making in Humans and in Artificial Intelligence,” Vol. 2, ed. I. Smith et al., Int. Institute of Advanced Studies in Systems Research and Cybernetics, 2003, pp. 12-17
11) Herbert, Brian, and Kevin J. Anderson. (2019) Dune. The Butlerian Jihad. Tor, 2019.