The Nightshirt Sightings, Portents, Forebodings, Suspicions

The recent NSA domestic spying scandal that shocked everyone is not really so shocking if you are the sort of person who likes to think about the possibilities (and pitfalls) of knowledge. We are now in the era of “big data,” which is changing the landscape not only of state surveillance but also science and health. Big computing power is enabling not only the gathering and storage but also the synthesis of exponentially greater amounts of information from scientific studies and clinical trials than ever before. In the halls of national research institutions, these new developments are being lauded (rightly) as heralding a new era of truly unprecedented scientific discovery.

Arguably, science itself is about to undergo a singularity, because the next step beyond big data collection is automating the very conduct of research to gather that data.

We all know how robots have or will soon take boring and dangerous tasks like vacuuming our floors or fighting our wars out of human hands. But no one thinks about the infinitely tedious task that is doing good science. In not too long, we will have the ability to automate not only the gathering and interpretation of information but also the very posing of research questions, and one of the first things we will teach intelligent computers to do is to ask questions in a scientific fashion—that is, form hypotheses based on prior findings, and then design and perhaps even (with the help of robots) conduct experiments to test them.

Whether artificial intelligence will ever become sentient (let alone spiritual) as Ray Kurzweil anticipates, AI will nevertheless be able to carry on the scientific endeavor increasingly independently, on a massive scale, fast, and without human biases and egos. This, whether we like it or not, will be truly objective science, which has never quite existed in the past even if it has always been the goal.

Space Probes Multiplying Like Rabbits

As long as we touch wood, Carl Sagan-fashion, and add “If we do not destroy ourselves,” the ability to automate science and deal with exponentially greater quantities of data will revolutionize our understanding of the natural world. Progress in medical research and many many other fields will make leaps and bounds, and it will ultimately revolutionize the exploration of space.

Because remember that, along with the big data and AI revolutions, we are also at the birth of the 3-D printing revolution. The use of local resources to create copies of machines and other supplies transmitted through space as simple information is going to make human life and work on the Moon and Mars and the asteroid belt feasible; and coupling a 3-D printer to a smart probe or drone will at last give us exactly what Von Neumann envisioned as the tool any advanced civilization will use to explore beyond its solar system.

Once a 3-D printer prints out another 3-D printer, the robot reproductive system is a reality. Interstellar probes thus equipped can replicate themselves at their destination and thus propogate from planet to planet, star system to star system, completely autonomously. Because they can perpetually repair themselves and reproduce, such probes would have limitless durability, and this would give them limitless patience. And what would limit them in the amount of reproduction that they do? Imagine: A thousand self-replicating probes are dispatched to the nearest star systems, where they copy themselves and establish a presence on every planet, if only to observe the geology and meteorology so long as nothing more interesting is going on, and send copies of themselves to further star systems, and so on. When they encounter really interesting stuff, like life or another intelligent species, they would swarm such a world with probes and dig in (quietly) for the long haul.

Such probes would have a built-in motive for curiosity and ability not merely to observe and record but to behave like experimenters: to generate their own hypotheses, design experiments to test them, and tediously replicate and re-replicate their findings alone or collaboratively, to constantly nuance and update their deepening understanding of their subject species. Such probes will not be passive, in other words, but will also interact in a very precise, deliberate, controlled fashion, and repeat these interactions obsessively in the same and different conditions, tirelessly, again and again and again, building up conclusions of high confidence. They will be more than “probes” as we usually think of them, but full, autonomous science platforms, continuously sharing data among themselves and constantly or periodically relaying that information to each other and back home for storage and future use by the civilization that built them (or their robot protectors).

Revisiting the ETH

Coupling the emerging reality of the Von Neumann probe with an expanding sense of just how “big” data could possibly get enables us to re-think some of the criticisms that have been leveled at the extraterrestrial hypothesis (ETH) in years past, because some are based on assumptions about the manner, scope, and aims of ET data collection that strike me as increasingly questionable.

Jacques Vallee was among the first ufologists to question the ETH. Vallee was in on the ground floor of the Internet and has been a technological as well as Fortean visionary throughout his career, and his argument against the ETH as an explanation for the full gamut of UFO contact experiences throughout history remains powerful and persuasive. I have come around to thinking he’s right—there’s a lot more to the UFO story (especially abductions and flying saucers) than flesh-and-blood aliens traveling across space to visit us. However, one of the five pillars of his argument—that the estimated millions of “landings” just in human history vastly exceeds what would be needed for a survey of our planet and civilization—misses what I suspect would be the very nature of any extraterrestrial “agenda” that was capable of populating space with swarms of robots and gathering and using big data over the long term.

In his 1989 paper “Five Arguments Against the Extraterrestrial Origin of Unidentified Flying Objects,” Vallee writes:

It should be kept in mind that the surface of the earth is clearly visible from space, unlike Venus or other planetary bodies shrouded in a dense atmosphere. Furthermore, we have been broadcasting information on all aspects of our various cultures in the form of radio for most of this century and in the form of television for the last 30 years, so that most of the parameters about our planet and our civilization can readily be acquired by unobtrusive, remote technical means. The collecting of physical samples would require landing but it could also be accomplished unobtrusively with a few carefully targeted missions of the type of our own Viking experiments on Mars. All these considerations appear to contradict the ETH.

Granted, it was 1989 when he wrote this. But besides predating the era of big data, this notion that ET space exploration would be satisfied with purely observational, “thin-slicing” data collection misses a whole side of science: experimentation and replication of findings. Remote, unobtrusive observation and periodic visits to collect samples would not remotely satisfy the scope of a full scientific research program on a potentially interesting planet such as our own. With massive data-gathering, storage, and synthesis capability wedded to machine self-replication, an “interesting” world such as ours could potentially have been swarmed with a hundred or a million probes, not only quietly observing and recording but also overtly interacting with the local flora and fauna for the purposes of experimentation and hypothesis-testing over the full course of its history.

Control Systems vs Psych Experiments

One of Vallee’s most far-reaching insights about UFO contact is that there is a regularity to it, a kind of ‘irregular regularity’ reminiscent of a reinforcement schedule in behavioral research. This insight supported his theory that UFOs may be some sort of control mechanism. The question is, control for what purpose? That UFO encounters represent an effort to shape our evolution is a popular view, and it could well be true in some cases. Yet the simple, scientific collection of behavioral data is another possibility that, despite what Vallee argued, is not at all inconsistent with either the the absurd, symbolic nature of UFO encounters or with their sheer number and repetition throughout recorded time.

First—and here I’m sure Vallee would agree—UFO encounters of all kinds (not just alien encounters or abduction experiences) not only resemble Zen koans but also resemble the contrived, surreal, occasionally uncanny situations that experiment participants find themselves in in any campus psychology laboratory. Even when they are aware they are part of an experiment, volunteers are generally deceived or not given full information about the purpose of the experiment. Experiments sometimes involve other “participants” who are actually confederates behaving in a realistic but specified manner in order to provoke some kind of response or decision on the part of the volunteer. Any experiment will also include at least two groups differing on a single parameter—a control and an experimental condition, in other words. Generally a single study will be part of a series, a whole research program, in which multiple experiments test numerous variations on a theme, in order to increasingly refine our understanding of a given psychological process.

Crucially, one of the keys to obtaining reliable, predictable data in psychology as in any other field of science is obtaining a large enough sample size. Thus, you recruit as many different volunteers as your grant money affords, and you run the experiment enough times that even a small behavioral difference between the conditions will achieve statistical significance and thus pass muster as a robust finding. Then, there is the need to repeat the experiment across different laboratories and replicate the finding so everyone can really trust it.

Repeatability of findings happens to be a huge problem across our sciences these days, since perverse reward incentives (tenure and grant competition, etc.) and other problems such as fraud are leading to the publication of data that are not as robust as they seem at first glance. But imagine those perverse incentives weren’t there. Imagine you were a “science machine” with all the time in the world and thus infinite patience, and with no pressure to publish or obtain tenure with startling findings, and your only goal was to acquire a truly “thick” understanding of how humans behave and react to specific circumstances with extraordinary confidence. Part of this imperative would include grasping that the species being studied is highly complex, that it is culturally and socially and psychologically adaptable and even biologically still evolving (and that your own actions may contribute, at least in a small way, to that evolution).

It would mean, I think, that you would endlessly devise new experiments to test new and different emerging hypothesis, run those experiments with large enough numbers of humans that your findings would be robust (but not so many that you ended up interfering in a significant way with the species as a whole); and it would mean that you would need to re-run the various experiments again and again and again, ceaselessly, throughout history. Many, many “landings,” in other words. As a science machine, you would be doing a lot of science, again and again, interacting just enough to test hypotheses in large enough samples, but not betraying your true purposes to the “volunteers.”

Thin Slicing vs. Deep Anthropology

As far-thinking as Vallee was (and is), he formulated his critique of the ETH at the toddler-hood of computer technology, so may have tended to think of the limits of information in human-experience rather than computer-experience terms—in other words, of isolated visits to reconnoiter and gather samples and “report back” somewhere. Carl Sagan envisioned extraterrestrial contact with earth in similar terms—periodic visits (every 10,000 years or so, he suggested). But when science can be undertaken completely by locally-based machines with storage capacities (locally or remotely) that vastly exceed even the computers of our NSA—and that are coordinated, unbiased, totally patient because they have nothing else to do and nowhere else to be, and moreover can collaborate in large numbers because they can reproduce themselves—then a level of science could probably be achieved that would be hard for any human to fathom.

One might ask why an extraterrestrial civilization would want to engage in such “deep anthropology,” but political and scientific realities of our own time make the answer pretty evident. Our scientific, technological society is already built on centuries of “basic” science—that is, science undertaken for its own sake, usually without any direct or foreseeable payoff in application. Knowing the mating habits of deep-sea protozoa may seem useless to most people (including many taxpayers who do not understand the importance of this kind of science), but scientists and smart policymakers who fund the science know that these details are all part of a big puzzle and any bit of information may ultimately pay off in unforeseen ways, years or decades or centuries down the long road. Thus, are our basic curiosity about the universe, and our social ability to invest resources in that curiosity, adaptive.

More basically, knowledge is power (or at least, security). It enables prediction and control. If money were no object, there is certainly no limit to the degree of prediction and control an intelligence agency like our NSA would like to achieve over even the remotest long-term threats to a nation’s security—for obvious reasons. As much as we may balk at the kind of surveillance our spying programs engage in, if it could be shown (and surely they will attempt to do so) that another 9/11 would be averted through such deep, extensive data collection, then some people would have no problem with the loss of privacy entailed. Likewise there is no limit to the degree of prediction and control—over illness, for example—that researchers at NIH would like to attain, given unlimited funds. If a cure for cancer can come of vast linkages of medical records and trial data, who will dispute such a project?

We thus need not even invoke any “anthopocentric” motive of pure curiosity to see why an alien intelligence or civilization will, when capable of “learning all that is knowable” inexpensively and in an automated fashion, embark on such a project. Such a civilization will have gotten to where it is through the same path we did—through science. When the kinds of constraints we now still face, in terms of funding and resources and the limits of human bias and the limits of processing power and storage, are overcome through advanced artificial intelligence and robotics, such a species/civilization will be in a position to undertake knowledge acquisition of mind-boggling scope and resolution, and it will have no reason not to.

That civilization will send its eyes and ears and roving brains outward, everywhere settling in for the long duration, in large semi-coordinated numbers, learning all that is learnable over the whole history of every star and moon and planet, about its geology and weather and even its primitive flora and fauna (if any)—because who knows what will happen in a million or a billion years? Who knows where life will emerge from primordial muck? Who knows what tree-dwelling mammal might become a spacefaring, militaristic civilization down the long road, and thus be worth settling in with and watching closely and learning how to predict and control should that species ever pose a threat to its security?

—

• Tweet