Friday, September 28, 2007

My little ROBO SURVEY

I'm running this by my students, and some of you may find it in your email inboxes soon enough. On the printed version there is more space for writing between the questions. I've already done nearly 200 of them, so it's too late to ask for ideas as to how I can change it. Even so, I'm very open to suggestions for future versions of this quiz, so, fire away!

(Below is a version of the poll that uses a more easily quantifiable response parameters, with integers ranging from -2 to 2, but it's still in rough draft form, so I welcome your suggestions.)

I may work the data up into a report on "Perceptions of AI in Taiwan", maybe something for The New Atlantis or some such. Incidentally, Taiwan seems to be as much into biotech as Japan is into robotics. If you'd like some tunes to groove to while perusing the survey(s), I highly recommend Volta's "Robot Needs Oil", as performed by a quartet of Sony's QRIO robots.

[One other important class project I want to work on is getting my students to write short letters to the President on behalf of the people suffering in Darfur.]

Oh, also, I did recently finish Rodney Brooks's Robot and Steve Grand's Growing Up with Lucy, both of which found refreshing for their frankness about the so to speak impressive failures of the last decades of AI research. Moreover, I fully agree with their emphasis on the importance of embodiedness and situatedness for AI systems (viz., relying on sensory processing more than abstract reasoning codes and using social interactions rather than idealized problem goals). They are of the opinion that the best way to approach AI is via AL (artificial life), which is to say that if you make a system biomimetically dynamic enough, it will eventually produce what we call a mind. Grand sees GOFAI (good ol' fashioned AI) as trying first to program a mind, as a set of logical operations, and then plug that software into hardware to make a brain, an approach Grand considers a dead end. Better, he says, is to first make a brain (by producing "virtual neurons" on with the convolution of numerous parallel processors), which will form its own inner parameters based on real interaction with a real environment. Depending what the assigned goal is, the numerous substructures will mutually alter each other so as to achieve that goal within a certain "yin/yang" action potential, much like a servo motor will adjust its own gears to maintain a certain acceleration, torque, or angle, or a thermostat will adjust the AC and heater to keep within a certain specified temperature range. Indeed, a hierarchy of servomotors could all collaborate to maintain a much higher-level goal, like a plane heading for a certain bearing and adjusting lift, rudder, speed, etc. all towards that one meta-goal on the horizon. The minded brain, Grand might say, is but one complex of biological servomotors that all collaborate to achieve larger goals, like getting food, sleep, etc. Hence, an artificial mind can only best come about by working up from simple neural patterns formed in contact with the world, along with higher programmed goals. As a connectionist SAIer, Grand wants, unlike a serial SAIer (i.e., mind first, then hardware brain), to make a brain first and let the mind emerge. He is honest time and again you can't say his Lucy really thinks or understands, but who cares, he asks. Why need we make a thinking robot when a responding robot will do just as well for its own given action potentials and goals?

Interestingly, Lucy started out as a model airplane (hence all the servo motors), and this only adds a poetic endorsement of my basic take on AI: robots are and always will be toys, tools, or works of art (and maybe a blend of those three). And anyone that thinks toys, tools, or works of art think, is just playing around with good thought. Just say that out loud, and pause to let it sink in: "Toys can think. Tools can think. Works of art can think." These are absurd claims once they are finally put in such clear terms, but such sobering clarity is exactly what an appraisal of AI requires. AI gizmos, like Aibo and Asimo, seem as real as they do precisely because our own human brains are designed to respond to coherent, dynamic, intentional behavior, which is to say, most of AI's public appeal is but a function of the human public's instinctive appeal for all things dynamic-within-limits. In this sense, AI is the most impressive optical illusion ever crafted by the very species designed to fall for it. It may take a bit of shaking to clear the anthropomorphic fog from your head when you see Asimo scamper here and there, but the fact is, behind his mysterious (and therefore mysteriously defensible) face plate is nothing but more circuitry. No thought. No feeling. No mind. No one there at all. He is a scampering shadow of the bright light of human reason, which is itself, a candle lit by the sun of divine wisdom.

In any case, I don't deny SAI because robots are said to think; I deny it for anterior philosophical reasons concerning the very nature of thought and language. I come to the table denying SAI (but not AI in the sense of animalistic response toys and non-intellectual goal-seeking devices) because I come to the table believing mental operations are inherently meta-physical actions, whether mediated by silicon or synapses. I don't claim only human brains can think precisely because I don't even claim human brains can think. Humans think, not things! Our brains do not think; we use our brains to think. The difference is as elementary and crucial as realizing that, on the one hand, we cannot see without our eyes, and, on the other, that our eyes do not see. We cannot walk without feet, but that does not mean our feet walk. And so for brains. I don't deny SAI for anthropocentric reasons, but because I reject the philosophy (or cluster of philosophies) on which SAI is predicated, namely, materialism, physicalism, epiphenomenalism, etc. I don't deny SAI because of technological limitations, but because of philosophical objections to the very idea of "constructing thought", an aim which immediately reveals a confused and impoverished understanding (oh the irony) of what intellection per se is. The immaterial and universal nature of thought precludes not simply the idea that machines could think but also, more fundamentally, the entire idea that material entities, like our brains, generate immaterial thought. An immaterial effect requires an immaterial cause. Thought is a specifically human function of the human soul, and it only makes sense that this capacity (as a formal power) is manifested in a suitably complex neural medium, the brain (as a material bases).

This is not to say the mind is some "ghost in the machine", for that view is Cartesianism. What I am arguing for is, as I wrote about earlier, classical Catholic (Artistotelico-Thomistic) hylemorphic anthropology (or ATA). ATA in no way denies the intimate and practically inseparable union of mind and body. It simply claims that the principle that gives coherence and vitality to the body, is the soul, and that, further, the highest power of the human soul is the intellect, a power which does not extract the soul from the body, but raises the mind to realities outside the strictly corporeal/sensory range (things like concepts beyond what is immediately perceivable and mental operations that definitionally overflow their physical instantiations, e.g., how many ways can you imagine completing the operation "p + q = r" or "If p, —» q. p, —» q"?). In ATA, the mind is a manifold reality. Intellect, as the power of abstraction to meta-physical and therefore supernatural reality, is the highest power of the soul, but hardly its only power. The soul expresses as well a sensitive (or sensory) mind, an appetitive (or emotional) mind, and a basic somatic (or homeostatic) mind, all or some of which are also expressed in the vast majority of living beings. ATA thus allows the soul to operate as a fully and continuously somatic mode of being -- we are not spirits in bags of skins, which strangely enough is what Minsky-style AI reduces us to: an omelet of abstract conceptual drives in a meat suitcase! -- but does not trap the mind within purely empirical, sensory data. Thus, I am free to agree with Grand and Minsky and Brooks, et al., on the mosaic-like nature of the majority of our neurological abilities, a mosaic that constitutes the bulk of our mental lives as basically automatic servo-neural networks in the symphony of the modular brain; and I can fully affirm the evolutionary "layering" evident in the brain as we move up from the basic reptile brain, to the woolly mammal brain, all the way up to the hominid cerebral cortex; and I can easily concur that without somatic input our mental life would be virtually nil -- even so, despite all this, I can in no wise agree that such neurobiological facts (or, I should in some cases say, conjectures) refute the existence of a power by which all such powers and modules are harmonized and, more importantly, by which the very concepts I'm discussing are grasped, analyzed, synthesized and articulated, namely by the intellect which elevates man, in material continuity with his somatic being, to contemplate all such ideas with transcendent self-awareness.

The soul, and thus the mind, is not floating "out there" outside the body -- it pervades the body as the wave pervades the undulating water or the impressed mold forms the wax. This is why all attempts at localizing consciousness -- whether in the pineal gland (à la Descartes), or in the frontal lobes (sort of à la Baars and his workspace of the mind), or in somatic sensory markers (à la Damasio), or in quantized microtubules (à la Penrose and Hameroff), or in synaptic bottlenecks (à la Grand and Edelman), or in any constellation of modular functions (à la Stanovich and Minsky and Brooks), or in genetic loops (à la Baum), or, indeed, in the nervous system itself (à la the reigning materialism of contemporary neuroscience) -- all such localization is a wild goose chase because the mind, being immaterial, is not localizable: pervading the "field of a person", the soul is everywhere and yet not in any place (in which case I am a noologial nullibilist [hat tip to Thomas More], but that's quite literally "neither here nor there", haha!). At every level, the evident mental power of a person can be traced up a level, not stopping at the rising (and ever hipper) tide marker of Heideggerian AI (à la Dreyfus), nor even at the imaginary heights of the self-as-narrative-fiction (à la Dennett), but rising infinitely to the intrinsic goal (nisus, entelechy) of human nature: perpetual and total relatedness in analogous glorification of the triune God and in mystical communion with Him. The mind is always shifting its position on the stage of cognitive research because it is always shifting its position closer to its final goal, the beatific vision, and the glorified transformation of -- not escape from! -- the body. The soul is sacramentally present in the body as Christ is sacramentally present in the Eucharist. Once the free person, then, freely encounters this hidden presence with its own hidden self, there is a union which not only draws the immaterial capacities and tendencies of the soul to contemplate thr divine, but also incarnationally draws the person into the social and ecclesial web of embodied existence.

Clearly, you can guess how I might answer my own surveys, but don't let that bias your replies heheh!


  • Results will remain anonymous. 結果會保密的。
  • You may continue writing on another sheet of paper, if needed. 如果在這紙上寫不下,就可以在另外一張繼續寫。
  • DATE 日期:
  • NAME 名字:
  • AGE 年齡:
  • NATIONALITY 國家:
  • OCCUPATION 職業:

1. What is a robot? 你覺得什麼是機器人?

2. What is a human? 人類是什麼呢?

3. At this time, are there things humans can do that robots cannot do now? 現在有什麼事情是人類可以做,但是機器人不行?

4. Are there things humans can do that robots will never be able to do? 有什麼事情是人類可以做,但是機器人絕對不可能做?

5. At this time, are there things robots can do that humans cannot do now? 現在有什麼事情是機器人可以做,但是人類不行?

6. Are there things robots can do that humans will never be able to do? 有什麼事情是機器人可以做,但是人類絕對不可能做?

7. Do you believe robots can or ever will be able to i) think (i.e., be conscious), ii) understand, iii) feel, iv) love and/or v) sin? Why or why not? 你相信機器人有一天能夠 i) 思考(有意識), ii) 理解 , iii) 感覺 , iv) 真愛/情感和 v) 罪惡感嗎? 為什麼?

8. Do you hope humans can someday make real artificial intelligence robots? Why? 你希望人類能夠發明人工智慧的機器人嗎? 為什麼?

9. Do you think robots will make human life better or will they be a threat to humanity? That is, will robots be useful pets/toys or will they take over humans? 你認為機器人會對人類有利或者造成威脅? 換句話說,你認為機器人會成為人類的寵物/玩具或者取代人類?

10. If you had a robot, what would you name it? 你擁有機器人的話,會幫它取什麼名字呢?

Ω Any additional comments? 還想說的話?






1 comment:

the Cogitator said...

Hi Michael,

I assume at some point you'll see this. I'd much prefer to talk about anything over beer! Thing is, I'm still trying to cut out some evening classes. Let me clear my evenings in October and then I should have more time to meet. You can email me at fidescogitactio AT gmail DOT com . Karl and I had a good talk a couple years ago, and we were set up to meet more regularly, but once I got back from World Youth Day in Köln, he did not, as far as I could tell, return my emails. Perhaps theology on tap was more his style and once the tap ran dry, his interest in dialogue evaporated as well. Are you both atheist buddies athirst?

Let me just say, though, that I find your frequent dismissal of philosophy is perturbing. The very purpose of philosophy is to ask questions about what questions science is asking, so merely brushing off dualism because cognitive researchers are doing so is highly, umm, woolly. Noology is not the proper study of cogsci, so what it purports to say about dualism is buried in grains of salt. Dualism is not established by findings in cogsci, so using the latter to discredit the former misses the point. It is not methodologically licit to discredit ATA with neuroscience, so how is it licit to ground mind-body identity and materialism in it? In any case, I am surprised you say cogsci is only starting to abandon dualism -- isn't that claim a century or two ex post facto?

Also, what exactly have you read by Fr. Jaki? Specifically, have you read **Brain, Mind, and Computers**?

You can email me, as I say, but silence will only reinforce your air of a troll at FCA.

What beer do you drink and where might you like to meet up?