The “Observer Effect”, Consciousness, and the Singularity: A Turing Test
In World War II, a brilliant computer scientist named Alan Turing broke the Nazi's encryption system known as "Enigma". His work helped win the war for the Allies, but his legacy extends far beyond it
This post was originally written on my WordPress site years ago, but I’m reproducing it here because I want to connect it to a set of related posts that I hope to weave together into a coherent whole—a “big idea” that I have been chasing for a while.
As an introduction: there is a computer-science concept called a “Turing Test”, which is named after the mathematician and cryptographer Alan Turing. He helped crack the encryption of the so-called Enigma machine that was used in World War II by the Nazi High Command.
The Enigma machine was used to encrypt command-and-control messages sent between Hitler’s Central Command staff and the many German military commanders fighting battles all over Europe and under the seas of the Atlantic.
Turing cracked the code, allowing the Allies to disrupt the Nazi’s war plans, and he invented what would become the modern “computer” in the process.
His life story is captured in the movie “The Imitation Game” and also in the book “Cryptonomicon” by Neal Stephenson.
While thinking about “computers” in general, Turing realized a time would come when a “machine” might possess a level of “intelligence” that could be indistinguishable from a human being. To try to tell the machine apart from the human being, he devised what we now call a “Turing Test.”
This is sometimes described using the following analogy: you put a computer, and a person, into separate rooms. You put a phone in each one, and an investigator is given the task of calling the phone in each room from the outside and asking any series of questions they like as they converse with the person or machine.
The goal is for the investigator to try to determine, using only the responses they hear to the questions they ask and the conversations they hold, which room has the computer in it, and which has the human being. If they cannot tell, then the machine is judged to be “sentient” in some sense.
Here is the original article I wrote, which introduces a surprising and unique idea involving a new way to perform a Turing Test using a consequence of Quantum Physics; it is an idea that has troubling implications.
I’m in the process of writing a “cosmology” themed book with a difficult and complex thesis: explicating the role that living things might play in the long-term evolution of the Universe, through the mechanism of entropy reversal.
Entropy is sometimes described as a measure of the “disorganization of matter.” The laws of Thermodynamics assert that it always increases in the Universe as a whole.
The book’s overall topic is far too deep to delve into here, but I came across some material recently that may find its way into my book, and the thinking I did around this new topic lit up some fresh ideas.
To thoroughly grasp this “essay”, you’ll have to watch both videos linked below, then come back and read this again.
On 12/10/2019, I watched this video:
This video covers a fascinating bit of research that purports to prove, in a fairly robust way, that human conscious thought can—statistically speaking, anyway—influence the outcome of a famous quantum mechanics experiment known as the “double slit” experiment (for details, watch the video.)
The experimental subjects in this study can—apparently—imagine themselves influencing a certain condition in the double-slit apparatus, causing something to be “observed” (thereby “collapsing the wave function”) such that this conscious effort apparently has a quantitatively measurable effect on the outcome of the experiment.
Their thoughts can change the behavior of a distant experimental apparatus.
While I have trouble acknowledging this (I’m a Quantum Mechanics heretic, thanks to my physicist friend Dr. Randell Mills, and I never really believed in things like telekinesis or parapsychology) I have no choice but to accept that these guys are evidently onto something with this experiment; the results are compelling.
I have objections to the idea that the current form of Quantum Mechanics (QM) —that is, the physics theory, from a mathematical/framework point of view—is the correct mapping of theory to the real-world physical phenomena of “quantum physics”.
There are many well-known “quirks”, gaps, and flaws involving the QM framework that are dissatisfying and reveal it to be arguably incomplete.
To be clearer: I acknowledge the validity of (at least some or most of) the phenomena that are described as consequences of quantum physics, even though I disagree that our current abstraction of how to represent these phenomena mathematically via QM is the “best fit”.
I think Dr. Mills work is a better fit in most cases (although this double slit result is hard to reconcile.)
In any case, the researchers behind this particular experiment seem to have been meticulous in removing all possible extraneous mechanisms that might interfere with their experimental results, and they conducted these experiments with fairly ingenious controls and delay protocols that make their conclusions very difficult to refute.
They were also thorough in their use of meta-analysis techniques to inspect their raw data for evidence of bias (p-hacking and selective reporting, for example.) There are also efforts underway to duplicate their results which appear to show confirmation.
The results seem to show, among other things, that “meditators” (people who practice meditative thinking) can achieve more statistically significant effects on influencing the double slit result than others; but humans, when “observing” the experiment have on the whole a statistically significant influence on the double slit result in contrast with simple machines.
The experimenters used an ingenious setup involving the internet as a distance separator and had a Linux “software bot” sometimes take the place of actual humans in performing the experiment, which involves interacting with a web page displaying some sort of feedback graph. This turns out to be a task that can be automated by a “bot” with a small bit of clever web programming.
The “Linux bots” – machine “observers” that sometimes took the place of actual humans — showed results that would be expected by pure random chance (i.e., no QM “observer effect”) whereas humans showed measurable QM “observer effect” results with (so the experimenters claim) 5-sigma statistical significance. This is a threshold that qualifies other experiments in physics as showing valid results.
Let’s assume for the moment that this experiment was conducted meticulously and actually shows something valid: that human conscious thought affects the outcome of a distant double-slit experiment, thus confirming the hypothesized “observer effect” of Quantum Mechanics. (This is a mind-blowing result all by itself, by the way.)
This led me to a few speculative questions of interest:
Do children show the same magnitude of effect on this experiment as adults? Does it vary by age? Is there an ‘age of consciousness’ metric that can be derived from doing this experiment on different age groups—meaning can you find an age at which this capability to influence the double slit first appears?
Can animals (dogs, dolphins, elephants, chimps, crows) be “taught” the feedback method used in this experiment well enough to serve as “conscious observers”? If they can, what does this mean?
If we re-define consciousness very narrowly as “that state of mental activity which has the known and measurable 5-sigma effect on this new double slit experiment” — if we define consciousness in this way, knowing that simple “Linux bots” do not show this effect—can we use this as a sort of “Turing test” for eventual “AGIs” (Artificial General Intelligence) to determine when they eventually “cross the threshold of cognition”?
In the video mentioned below, one of the speakers relates some cognition experiments with babies, wherein at a certain age (8 months) they become “aware” that gravity has an expected manifestation.
Eight-month-old babies, when presented with hovering objects that seemingly defy gravity, show signs of awareness that “something is amiss” (whereas 5-month-old babies do not.) At this age, babies are also aware of the concept of “self” via mirror experiments. They see themselves in a mirror and understand “me”.
Again: assuming that it is a valid experiment, can this “double slit” experiment—or one like it—be used to narrowly determine the “age of conscious awareness” as children mature?
Does it correlate with this 8-months old cognitive state? If so, what accounts for what happens to children’s mental abilities when they cross this “threshold” and are able thereafter to influence a double slit outcome? Does it correlate somehow to neural connectedness or density?
On 12/11/2019, I then watched this video on AI:
Having just watched the “double slit” video the day before, I immediately grasped the applicability of the double-slit experiment as a kind of “Turing Test” to determine whether/when an AGI achieves “conscious awareness”.
If an advanced machine (much more sophisticated than the “Linux bots” of the prior video) can eventually “influence the double slit” outcome in the same way (or better!) than humans can, perhaps this serves a “proof of consciousness”!
A positive Turing Test result.
This may be another, more effective discriminator to determine when the “singularity” event horizon has finally been crossed, per Ray Kurzweil’s characterization of this threshold as the point in time when machine intelligence becomes “sentient”.
From the connection of this experiment to some of the topics covered in the AI video, I speculate as follows: we have already created artificial intelligence systems that significantly outperform humans in chess, the game of Go, image classification, etc. As improvements continue (and borrowing from Kurzweil’s perspectives on this) there will come a time when a generalized artificial intelligence “crosses some threshold” (i.e., the “singularity”.)
If machine-based life evolves to demonstrate “consciousness” and then is further able to exert this QM observer effect, and if exploiting this effect, it is orders of magnitude more capable at this than modern humans… oh, my.
That will have unimaginable consequences on the state of the Universe, “post-singularity.” Because this means “spooky action at a distance” could be exploited by artificial life perhaps for superluminal communication (or God knows what else.)
Another interesting topic in the above video surfaced when some panelists were discussing the “purpose of consciousness” and the purpose and meaning of human emotion as an origin of moral force and as a motivating impetus of human activity in general.
Restating things in my own words: emotions are a stimulus/reward system that exists to provide an “impulse” for humans (through consciousness) to do whatever it is that we do. Presumably, machines will never possess this.
How I tie this into my book thesis is as follows: I believe that humans (and other living things) are “designed” by our Creator to fulfill a particular role in the evolution of the Universe—the long-term reversal of entropy (just why this is so is the topic of the book.)
Living things do so by creating lower-entropy “ordered states of matter”: the ultimate example of which will be the highly ordered states of matter—advanced artificial life—that arises when machines become self-aware after the singularity, which will then reproduce exponentially per Ray Kurzweil’s vision.
Consciousness and emotional drive are the mechanisms that “compel” living things (particularly humans) to act out their role as entropy reducers. These drives represent the wellspring of humanity’s creative force.
Consciousness appears to have a direct relationship to states of matter (via the QM observer effect.) Post-singularity “artificial lifeforms” might, therefore, “evolve” something analogous to emotions, in order to provide the “motivation” for their own continued expansion throughout the Cosmos. Or perhaps not….
If they did not do so, what “motivation” would “compel” these new life forms to continue evolving and expanding?
Another interesting talk about consciousness which connects some of these threads together:
update: Add this to the mix. If this is already possible, what might an advanced AI do with this?
update: The end of this video also touches on these topics: animal consciousness, children, and AI
More:
Have been thinking about this area a lot recently. I began wondering whether there was a difference between something that was merely unknown, and something that was unknowable?
In his 1950 paper, “COMPUTING MACHINERY AND INTELLIGENCE”, Turing describes the Laplace Demon with reference to digital computers. The point being, if the state of universe is known (or perhaps in principle knowable), and everything is working according to rigorous mechanical laws, then we have no free-will. Computers, with finite discrete states, are in principle knowable, and the results of their calculations (saying running an NN) are knowable in advance (by a faster computer).
In the real world, however, we are granted “adequate determinism” only, as chaos theory and QM ensure that we always surrounded by unknowability (my spell checker didn’t know that word for example).
You say that, according to thermodynamics, the amount of entropy in the universe is always increasing. If you don’t know already, you might be interested to check out of the work of David Layzer (circa 1975), who showed that the amount of “information” in the universe is also increasing, despite the increase in entropy. This in itself should be the final nail in coffin of determinism. But more importantly, where is this “information” coming from?
I’m not sure about this “singularity” thing. It’s talked about as if it’s an established fact, but has never been observed in nature. I’m kinda of the view that if it were going to happen, some other civilisation would have unleashed it on us by now. The universe seems to go to lengths to deny us full knowability and omniscience (instead we get free-will). Perhaps it also guards against other things too — for example, space is expanding such that we could never reach distant galaxies even if we set out today at the speed of light. Therefore, some “machine singularity” in a distant galaxy may never reach us, even it wanted to. It will just have to shake it’s digital fist at God!
Personally, I just want a cleaning robot, that has free-will, enjoys its task, and is glad to see me when I come home. I have a feeling that such a thing must also be capable of suffering if it is be conscious in any meaningful way. No doubt I would cry over it if it ever broke down.
However, I can’t see any corporation intentionally building a “machine” with the free-will any time soon. Rather, they would prefer to lock into a digital Metaverse in order that we be sucked dry of information generating ability.
Great piece. We are on similar lines!
Thank you for a most interesting article.
I would like to acknowledge the Polish mathematicians who first broke the Enigma codes.
"Thanks to the efforts of the Cypher Bureau, the Polish knew 95% of the Germans’ order of battle before the invasion of Poland on the 1st of September 1939."
"With the storm clouds rapidly gathering, it became obvious to the bureau that a decision needed to be taken before their country was invaded and its work discovered. It was decided that the bureau’s work, its replica Enigmas and its bombes should be passed on to the British and the French. A meeting was arranged in the Kabaty woods outside of Warsaw on the 26th of July 1939. There, Cypher Bureau chiefs Gwido Langer, Maksymilian Ciężki and Rejewski and his team handed over everything the bureau had on their work cracking the Enigma machine to French cypher chief Gustave Bertrand and his British counterpart, Alastair Denniston. Denniston and the head British cryptanalyst, Dilly Knox, were stunned when they discovered just how advanced Polish codebreaking was. So far, the British had relied on linguists to try to crack Enigma messages. The Poles had proven that the key to cracking the code lay not in linguistics but mathematics."
Today there is a plaque at Bletchley Park which acknowledges the work of the Polish men. I first became aware of the astonishing work of this group of Polish mathematicians on a visit to the Polish city of Poznan, a few years ago.
https://www.history.co.uk/articles/the-polish-cryptographers-who-cracked-the-enigma-code