Brains need combinatorial representations because there are infinitely many possible linguistic representations, and the brain only has a finite storage space. Even a superintelligent system would benefit from combinatorial representations. Although a superintelligent system could have computational resources that are so vast that it is mostly capable of pairing up utterances or inscriptions with a stored sentence, it would be unlikely that it would trade away such a marvelous innovation of biological brains. If it did, it would be less efficient, since there is the potential of a sentence not being in its storage, which must be finite.
(iv) BISAs may have one or more global workspaces. When you search for a fact or concentrate on something, your brain grants that sensory or cognitive content access to a “global workspace” where the information is broadcast to attentional and working memory systems for more concentrated processing, as well as to the massively parallel channels in the brain [14]. The global workspace operates as a singular place where important information from the senses is considered in tandem, so that the creature can make all-things-considered judgments and act intelligently, in light of all the facts at its disposal. In general, it would be inefficient to have a sense or cognitive capacity that was not integrated with the others, because the information from this sense or cognitive capacity would be unable to figure in predictions and plans based on an assessment of all the available information.
(v) A BISA’s mental processing can be understood via functional decomposition. As complex as alien superintelligence may be, humans may be able to use the method of functional decomposition as an approach to understanding it. A key feature of computational approaches to the brain is that cognitive and perceptual capacities are understood by decomposing the particular capacity into their causally organized parts, which themselves can be understood in terms of the causal organization of their parts. This is the aforementioned “method of functional decomposition” and it is a key explanatory method in cognitive science. It is difficult to envision a complex thinking machine not having a program consisting of causally interrelated elements each of which consists in causally organized elements.
All this being said, superintelligent beings are by definition beings that are superior to humans in every domain. While a creature can have superior processing that still basically makes sense to us, it may be that a given superintelligence is so advanced that we cannot understand any of its computations whatsoever. It may be that any truly advanced civilization will have technologies that will be indistinguishable from magic, as Arthur C. Clarke once suggested [15]. I obviously speak to the scenario in which the SAI’s processing makes some sense to us, one in which developments from cognitive science yield a glimmer of understanding into the complex mental lives of certain BISAs.
SOME ISSUES FOR FURTHER REFLECTION
In the spirit of encouraging future discussion, I will close by raising issues for future reflection.
Given the vast variety of possible intelligences, it is an intriguing question to ask whether creatures with different sensory modalities may have the same kind of thoughts or think in a similar ways as humans. There is a debate in the field of philosophy of mind that is relevant to this question. Contemporary neo-empiricists, such as the philosopher Jesse Prinz, have argued that all concepts are modality specific, being couched in a particular sensory format, such as vision [16]. If he’s correct, it may be difficult to understand the thinking of creatures with vastly different sensory experiences than us. But I am skeptical. For instance, consider my aforementioned comment on viewpoint invariant representations. At a higher level of processing, information seems to become less viewpoint dependent. Similarly, it becomes less modality specific, as with the processing in the human brain, as it ascends from particular sensory modalities to the brain’s association areas and into working memory and attention, where it is in a more neutral format.
But these issues are subtle and deserve a lengthier treatment. I pursued issues related to this topic in my monograph, The Language of Thought, which looked at whether thinking is independent of the kind of perceptual modalities humans have and is also prior to the kind of language we speak [12]. In the context of alien life or SAI, an intriguing question is the following: If there is an inner mental language that is independent of sensory modalities, having the aforementioned combinatorial structure, would this be some sort of common ground, should we encounter other advanced intelligences? (Many of these issues apply to the case of intelligent biological alien life as well, and could also be helpful in the context of the development of SAI on Earth.)
The ethical and metaphysical issues surrounding postbiological intelligence concern me greatly. Perhaps the best way to introduce the ethical and metaphysical issues is to consider that the post-biological cosmos approach involves a shift in our usual perspective about intelligent life in the universe. Normally, we think of encountering alien intelligence as encountering creatures with radically different biological features and sensory experiences. The shift of focus is twofold: first, the focus moves away from biology to superintelligent AI, and this will involve theorizing about the computational abilities of advanced artificial intelligence. Second, as we reflect on the nature of postbiological intelligence, we must be keenly aware that we may be reflecting upon the nature of our own descendants as well as aliens. In essence, the line between “us” and “them” blurs, and our focus moves away from biology to the difficult task of understanding the computations and behaviors of creatures that will be far more advanced than we are.
What does this all mean? In contrast to Ray Kurzweil’s utopian enthusiasm for the singularity, I do not see normative evaluations of whether a post-biological existence is desirable for our species in the astrobiology literature, and there has been little discussion of the singularity within contemporary metaphysics and philosophy of mind. But it is important to reflect upon the ethical, philosophical and social implications of all this. Would superintelligent AI, including our own postbiological descendants, be selves or persons? Could they be conscious? My own view is that the question of whether AI could be conscious is key – if the synthetic being in question is not capable of consciousness, that is, if it doesn’t feel like anything to be it, then why would it be a self or person? I’ve discussed the issue of consciousness elsewhere [4], but since that point, I have been increasingly convinced that the question of machine consciousness is an open question that cannot be solved today. In addition to the matter of whether the substrate in question (e.g., graphene, silicon) supports consciousness, the devil is in the details of the particular AI design. That is, we would have to determine whether the architecture of the particular AI in question even employs conscious thought. Consciousness is associated with slower, more deliberative processing in humans, and it is unclear whether superintelligence would even need conscious processing, as it would have mastered so much already. What would be novel to it? And would consciousness even be associated with slower, deliberative processing in an AI in any case?
The science fiction treatment of androids may lead us to believe that machines can feel – for instance, consider the Samantha program in the film Her, or consider Asimov’s robot stories. But this is just science fiction, and the empirical and philosophical question of whether AI can be conscious remains open.
CONCLUSION
In this brief piece, I’ve discussed why it is likely that the alien civilizations we encounter will be forms of superintelligent AI (or “SAI”). I then turned to the difficult question of how such creatures might think. I provisionally attempted to identify some goals and cognitive capacities likely to be possessed by superintelligent beings. I discuss Nick Bostrom’s recent book on superintelligence, which focuses on the genesis of SAI on Earth; as it happens, many of Bostrom’s observations were informative in the present context [7]. Finally, I isolated a specific type of superintelligence that is of particular import in the context of alien superintelligence, biologically-inspired superintelligences (“BISAs”). I urged that if any superintelligences we encounter are BISAs, certain work in computational neuroscience, cognitive neuroscience and philosophy of mind may provide resources for at least a rough understanding of the computations of BISAs.
REFERENCES
[1] Cirkovic, M. and Bradbury, R. 2006, “Galactic Gradients, Postbiological Evolution and the Apparent Failure of SETI,” New Astronomy 11, pp. 628–639
[2] Dick, S. 2013, “Bringing Culture to Cosmos: the Postbiological Universe,” Cosmos and Culture: Cultural Evolution in a Cosmic Context, S. J. Dick and M. Lupisella eds., Washington, DC: NASA, online at http://history.nasa.gov/SP-4802.pdf
[3] Shostak, S. 2009, Confessions of an Alien Hunter, National Geographic (Washington, DC)
[4] Schneider, S. 2015, “Alien Minds,” in Discovery, Steven Dick, ed., Cambridge University Press (Cambridge)
[5] Davies, P. 2010, The Eerie Silence, Houghton Mifflin Harcourt (London)
[6] Bradbury, R., Cirkovic, M., and Dvorsky, G. 2011, “Dysonian Approach to SETI: A Fruitful Middle Ground?” Journal of the British Interplanetary Society, 64, pp. 156–165
[7] Bostrom, N. 2014, Superintelligence: Paths, Dangers, Strategies, Oxford University Press (Oxford)
[8] Guerini, Federico 2014, “DARPA’s ElectRx Project: Self-Healing Bodies Through Targeted Stimulation Of The Nerves,” http://www.forbes.com/sites/federicoguerrini/2014/08/29/darpas-electrx-p... Forbes Magazine, 8/29/2014. Extracted Sept. 30, 2014
[9] Miller, R. 1956, “The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information,” The Psychological Review, 63, pp. 81–97
[10] Schneider, S. 2011a, “Mindscan: Transcending and Enhancing the Brain,” Neuroscience and Neuroethics: Issues At the Intersection of Mind, Meanings and Morality, J. Giordano ed., Cambridge University Press (Cambridge)
[11] Seung, S. 2012, Connectome: How the Brain’s Wiring Makes Us Who We Are, Houghton Mifflin Harcourt (Boston)
[12] Hawkins, J. and Blakeslee, S. 2004, On Intelligence: How a New Understanding of the Brain will Lead to the Creation of Truly Intelligent Machine, Times Books (New York)
[13] Schneider, S. 2011b, The Language of Thought: a New Philosophical Direction, MIT Press (Boston)
[14] Baars, B. 2008, “The Global Workspace Theory of Consciousness,” The Blackwell Companion to Consciousness, M. Velmans and S. Schneider eds.,Wiley-Blackwell (Boston), pp. 236-247
[15] Clarke, A. 1962, Profiles of the Future: An Inquiry into the Limits of the Possible, Harper and Row (New York)
[16] Prinz, J. 2004, Furnishing the Mind: Concepts and their Perceptual Basis, MIT Press (Boston)
[Go to Top]
THINKING OUTSIDE THE SETI BOX
Seth Shostak
SETI Institute
189 Bernardo Ave.
Mountain View, CA 94043
seth@seti.org
Introduction
We consider the biological provincialism of traditional SETI, and why there are good arguments for thinking that the bulk of the intelligence in the cosmos is synthetic. Given this possibility, the SETI community should consider how to conduct a meaningful search for intelligence that is not constrained to habitable worlds. To that end, we consider some of the factors that might govern the behavior of highly advanced, cognitive machinery and some strategies that might aid in the discovery of same.
THE ANTHROPOCENTRIC BIAS
The premise of most SETI experiments, the Search for Extraterrestrial Intelligence, was established with Frank Drake’s pioneering Project Ozma more than five decades ago [1]. Today’s efforts differ in scale, but not in approach: Their strategy is to seek signals produced by cosmic inhabitants whose level of technology is at least as advanced as our own.
For more than two decades, SETI has been largely underwritten by private donations, and because of this the scientists involved are often pressured to make some estimate of the chances of success. To this end, they will frequently invoke the well-known Drake Equation which quantifies the number of galactic societies currently producing detectable signals. If some estimate of the prevalence of transmitting sources can be made, then a timescale for SETI success can also be made.
Unfortunately, the value of many of the parameters of this equation are still unknown, and the few for which new data have recently become available are little changed from the estimates made when the equation was first written. The Drake Equation, while ubiquitous and helpful in formulating the problem of SETI, does little to determine the odds for any particular experiment.
Of possibly greater importance is the Equation’s influence in setting strategy. It assumes that SETI will succeed only if there are at least a few thousand technically accomplished civilizations resident in the Milky Way. Detectable societies are assumed to consist of a large number of individuals, resident on a planet that’s not only amenable to life but also able to beget and sustain complex organisms. In other words, a world analogous to our own.
That view hasn’t changed in a half century. New thinking on how to conduct SETI has been less about the nature of the beings we seek or their habitat, and more about their presumed behavior.
As example, a matter of popular discussion is whether signals from extraterrestrials are more likely to be deliberate beacons, or accidental leakage. This discussion is largely motivated by the trend in our own society to shift to higher efficiency communication modes (e.g., direct satellites and fiber optics in place of traditional broadcasting.) This change has led many to opine that advanced civilizations will be economical, and not generate significant leakage. However, while this argument sounds plausible, there’s no denying that it is highly parochial, and based on human experience a scant century after the invention of practical radio and lasers. And even this modest speculation on the conduct of extraterrestrials – they will be more efficient users of energy than we are – has had little impact on SETI experiments.
In fact, experiments do what they are able, and are mostly indifferent to whether the signal being sought is intentional or otherwise. SETI today continues to adopt the playbooks of the past: the aliens are analogous to us, only more advanced. The circumstances of their environment are also presumed to be similar to ours.
Unsurprisingly then, SETI practitioners have been heartened by recent discoveries of exoplanets. The good news is that worlds akin to our own could exist in great abundance. Current estimates are that between 0.1 and 0.2 of all star systems host an Earth-size planet in the habitable zone [2]. This implies that tens of billions of these favored locales pepper the Galaxy.
But there is also bad news. At a time when the prospects for beings comparable to ourselves are improving, there is a slow-growing realization that biological intelligence may be only a short-lived – and possibly cryptic – stepping stone to the real thinkers of the cosmos: synthetic intelligence.
PROSPECTS FOR SYNTHETIC INTELLIGENCE
If researchers in the field of artificial intelligence (AI) are to be believed, we will invent machines that are our cognitive equals by mid-century. Roboticist Hans Moravec has pointed out that the exponential improvement in digital electronics will produce workaday computers with reckoning power comparable to a human brain in less than a decade’s time [3]. This rapid betterment in computation has led some, such as Vernor Vinge and Ray Kurzweil, to predict a future time – the “singularity” – at which our own intellectual capacities will be swamped by that of our devices [4],[5].
Of course, there are already machines that can outperform the human brain in tasks generally regarded as “intelligent.” The best chess playing computer can beat the best grand master, and the recent triumph of IBM’s Watson computer against seasoned contestants on a television quiz show attracted widespread attention, if not admiration. More recently, Google’s AlphaGo software beat a world expert human at the game of Go, one that is considerably more complex than chess. But as AI entrepreneur Peter Voss has noted, these attainments merely point up the current situation in which one can either build a machine that is excellent at a narrowly scoped task (e.g., chess) or one that is quite mediocre at many things [6]. In order to challenge the intellectual abilities of humans, what’s required is what is termed GAI – generalized artificial intelligence.
It is not the intent of this essay to either review or critique developments in AI research, but rather to assume that GAG machines will appear – if not in this century, then in the next. The timing is of little consequence to the implications for SETI. But the events following this development are straightforward:
1. If our own example can be taken as typical, then GAI quickly follows on the heels of radio technology – within a few centuries.
2. There is no reason to believe that the evolution of “wet ware” – augmentations of our own brains – can keep pace with GAI.
3. Because artificial intelligence can quickly evolve (by its own design), it will soon outstrip the cognitive capability of biological beings.
4. Artificial intelligence will be self-repairing, and therefore of indefinite lifetime.
5. GAI will be the dominant form of intelligence for any society that has progressed even slightly beyond the point of being able to send signals into space.
6. Unlike biology, which has been “engineered” bottom-up, GAI will be engineered top-down. We cannot hope to forecast what talents or interests it will have, but the one aspect of its functionality that seems safe to assume is survival. This sounds Darwinian, and therefore biological, but is essential if we are to find GAI now, billions of years into the history of the cosmos.
The bottom line is simple, if disquieting: biological brains will beget synthetic ones. If this technical evolution is commonplace, then there’s reason to expect that the majority of the intelligence in the universe is non-biological. This intelligence would not be dependent on water worlds, atmospheres, or planets at all. Consequently the premise of most SETI – that we should expect to find signals from old, habitable worlds – could be wide of the mark [7],[8].
It seems probable that the future of our hunt for extraterrestrials will require more than just new equipment. We’ll need to rethink what it is we seek.
SO HOW DO WE FIND IT?
Adapting our SETI strategies to the challenge of uncovering GAI may sound simple at first. Nothing more is required than to put less emphasis on targeting habitable planets, or even individual stars, and simply scan as much of the sky as possible. However, there may be opportunities to increase our chances of success by augmenting this simple, brute-force approach with insights about the likely nature or behavior of synthetic intelligence.
First, we are probably well advised to avoid hubris. There may be little we can fathom about the nature of artificial intelligence that might be the result of millions of generations of self-improvement – improvement not predicated on the slight and random modifications of Darwin, but directed changes. Such intelligence will surely be as superior to us as we are to the nematodes in the garden. Consequently, we should not feel too sure about our speculations as to what AGI might do or how it might be detected. Imaginative ideas about the interests and activities of synthetic beings are plentiful in fiction, but these ideas are vulnerable to anthropocentric bias.
However, there are at least a few aspects of GAI that seem less suspect:
1. Assuming that for such machines more computation is better, they can be expected to prefer locations with abundant energy and an effective heat sink. The former suggests the neighborhoods of early-type stars or black holes (either of the stellar variety or the massive objects hunkered at the centers of galaxies.) It’s been suggested that the outer regions of galaxies might be preferred locales for such machines because of their slightly lower temperatures, resulting in greater thermal efficiency [9]. However, given that the efficiency depends only on a temperature ratio between source and sink, this argument is of significance only if the energy source is no more than a few hundred degrees, as space is cold almost everywhere.
2. The short timescales for self-improvement may set up a “winner take all” situation. Whatever machine first appears in a given part of the cosmos could endlessly trump others that arise, since even a cosmically short period of time is a great number of GAI generations, and the new kids on the block could never catch up.
3. Given the dangers present in the universe, a machine might wish to buy insurance in the form of backup machines. These could be kept at a distance that would minimize simultaneous annihilation, but linked to the mother machine so that updates could be continually offered. Detecting this telemetry might offer a way to discover GAI, although one can assume that the communication would be point to point and unlikely to be intercepted with our instruments.
4. Another possible organization scheme for GAI might be hierarchical. Social systems might make sense if the increase of information in a machine eventually becomes small compared to the timescale for interaction with other machines (the light travel time between them). In other words, if the new capability acquired per year by a GAI eventually becomes a very small fraction of the previously accumulated capability, then interchanging information makes sense, since that information is not rendered obsolete and irrelevant in the time it takes to effect the exchange.
5. Whether intelligent machines would have any interest in broadcasting (as opposed to point-to-point telemetry) is impossible to know. One metric for intelligence is the ability to foresee danger and avoid it. The cleverest GAI, by this measure, might be less concerned about revealing their presence with easily found signals. They might also wish to communicate with other such machines that are largely outside their light cone, as these would have information that they could not obtain otherwise [10].
These considerations offer a few plausible arguments as to where we should look for GAI. However, they promise little in terms of assuring SETI scientists that such machines would have any motive to make themselves known.
In the case of biological beings, we can safely assume the presence of curiosity, as this trait is necessary to divine the laws of nature and build transmitters we could find. But artificial sentience might not share this type of curiosity. Maybe after solving all the puzzles of science, GAI would be happy to indulge itself with endless entertainments – perhaps with Bostrom-like simulations [11]. If they are capable of self-repair (an assumption in all of the above), then it may be that their primary project is to forestall the heat death of the universe and an end to their own existence.
CONCLUSIONS
What might SETI practitioners do to increase their chances of detecting what is likely to be the most prevalent form of intelligence in the cosmos? Unfortunately, the list is short.
A search for unusual phenomena in the vicinity of high-density energy sources is a straightforward desideratum. Another is to consider that the oldest of such machines might wish to contact their peers in other parts of the cosmos to compare notes and offer novel information. This suggests an experiment in which SETI searches for signals (radio or optical) in the direction of stellar black holes or quasars that are antipodal. E.g., two stellar black holes on opposite sides of the sky might conceivably host AGI whose beamed data would pass through our neighborhood.
Perhaps the best strategy to find the universe’s intellectual giants is the least deliberate: simply be careful to note any unusual phenomena uncovered in the course of astronomical research. Are there nebulae with anomalous, depleted deuterium? Do some stars or galaxies display unnatural infrared excess, a possible tipoff to energy-intensive residents [12],[13]? Are there cosmological behaviors without natural explanation?
It is easy to design an experiment to find the aliens of sci-fi, for these are robustly similar to ourselves. But when you don’t know your prey, the hunt can be hard.
REFERENCES
[1] Drake, F. 1960, “How can we detect radio transmissions from distant planetary systems,”Sky and Telescope 39, 140
[2] Petigura, E. A., Howard, A. W., and Marcy, G. W. 2013, “Prevalence of Earth-size planets orbiting Sun-like stars,” PNAS, 110, No. 48, 19273
[3] Moravec, Hans 2000, Robot: Mere Machine to Transcendent Mind, Oxford
University Press (Oxford)
[4] Vinge, V. 1993 “The coming technological singularity,” Vision-21:
Interdisciplinary Science & Engineering in the Era of CyberSpace, proceedings of a Symposium
held at NASA Lewis Research Center (NASA Conference Publication CP-10129)
[5] Kurzweil, Ray 2005, The Singularity is Near, Viking Penguin (New York)
[6] Voss, Peter 2015, http://www.agi-3.com/technology.html
[7] Shostak, S. 1998, Sharing the Universe, Berkeley Hills Books (Berkeley)
[8] Shostak, S. 2011, “Seeking intelligence far beyond our own,” International Astronautics Congress, IAC-11.A4.2.4
[9] Cirkovic, M. M. and Bradbury, R.J. 2006, “Galactic gradients, postbiological evolution, and the apparent failure of SETI,” New Astronomy, 11, 628
[10] Windell, Alex Noholoa 2015, private communication
[11] Bostrom, N. 2003, Philosophical Quarterly, 53 No. 211, 243
[12] Carrigan, R. 2009, “The IRAS-based whole-sky upper limit on Dyson spheres,” Ap. J.698 2075
[13] Griffith, R. L., Wright, J. T., Maldonado, J., Povich, M. S., Sigurdsson, S., Mullan, B. 2015, “The Ĝ Infrared Search for Extraterrestrial Civilizations with Large Energy Supplies. III. The Reddest Extended Sources in WISE,” arXiv:1504.03418 [astro-ph.GA]
{ https://www.seti.org/ }
The Search for Extraterrestrial Life and Post-Biological Intelligence - PART II
![Evidence of Aliens: Triangular UFO spotted in Google Maps [+NASA footage]](https://i2.wp.com/tecake.in/wp-content/uploads/2017/11/UFO-Spotted-in-NASA-Footage.jpg?resize=350%2C200&ssl=1)
@MikeWehner
