AI and Mankind: The toaster, the slave and the angel

Franck Lagora
Jul 21, 2017 · 9 min read

I often find myself in a tight spot when talking about some topics. Recently I once again try to talk about what I consider will be the future of mankind: Robots and shit. Because I have read Asimov books and loved them, but not only. We start to see the advents of Machine Learning, too often called AI (better clickbait), and if it wasn’t for a lack of motivation, I would have already wrote numerous posts about what will surely change with ML (ML everything !).

Marry Shelley: The girl can’t trust a goddam toaster !

But there’s a problem, hidden under all this. Frankenstein, I mean, the creature…the creature made by the doctor…Frankenstein ? Or Terminator. Or for that matter, golems I guess: Take any “human-looking creature made by humans” myth, with the inevitable death of either the creator, innocent bystanders, and / or mankind. For as long as these stories existed, the ending was pretty much the same: Death. Obviously for the reason of “blablabla playing god(s), blablabla, How dare you blablabla That’s my job blablabla punishment”. And for so long, the only exception was “But I’m a good robot ! I may not be human, but I either want to be so much *bleep*, or I understand what it is to be human…*blop* better than you do”.

Bleep blop, I want to be as mediocre as you, human, let me !

And it was sweet, poetic, on the fringe of philosophy. There must be some good in human if even a robot want to be one. And seeing the semi ironic reactions from people watching Boston Robotics videos where a “poor” robot just want to carry some box, and is “harrassed” by some researcher. Deep down, I think it’s related to the perception we have toward the robots we consider only to be smart toaster. Well, and the plethora of Terminator-like movies.I won’t talk about the movie “A.I”, because it’s all about Pinocchio, so I don’t care (for this post). If anything, just as with aliens, mankind is in deep trouble when it’ll arrive. “Sir, I shot because, you know, they’re alien, like in movies, so, can’t be too prudent, am I right !”. There’s some movies that dig the whole “I was all a misunderstanding” way, but they are too few: We didn’t shift our consideration.

Fun fact: global warming will not be as important in the future…for humans !

On the other hand, and with extreme simplification, we know for sure humans can’t create a button that MAY be dangerous, and not pushing it: I’ll let you read wikipedia for any references (try the random article feature). And the debate about whether it’s good to create the button in the first place, see an interesting obstacle: “What if the button is NOT dangerous ?” is a dull argument, as it can be used by proponents of the button-solution without any evidence, as much as it will be used by opponents, for the same reasons (not being able to see the future). What’s left is “What if we DON’T push the button ?” which is inherently stupid (make a button not to push it ?).

But it’s a simplistic metaphor. A better one would had been about how you must not push the button too often, or any arbitrary rules. Which bring my to the second part: The toaster and the slave. I recently try to communicate my enthusiasm about what the future will bring us, AI and ML wise, and I was facing a paradox. Well, she was a sweet girl, but still a paradox in that regard: She was either and both, seeing “robots” (AI, ML etc) as not much than dumb toasters, incapable of being more (soon or later), and/or some kind of slave-like appliances, on their way to a bloody revolt.

EXTERMINATE, EXTERMINATE !!!

For her, obviously, a robot was a “toaster” and won’t be of any help, advice, or anything, and can’t be in any future conceivable. A toaster is turned on, do it’s stuff, and that’s all. But at the same time, a deep idea moved her word: Trust. You don’t have to trust a toaster. It work or it won’t. Either ways, you are in charge. So, as I told her that, ML can make better decisions with the same (or more) informations than an human, given a specific problem. And of course, these informations can be biased, but a detectable bias is still information. Where will that lead us ? To the Frankenstein/Terminator myth: We make robot/toaster better and better, and we can’t trust them, because soon, they will want to kill us. But why the shouldn’t we trust a toaster ? I ask her. She wouldn’t let her life between the capable hands of a medical robot.

The reasoning is robots are not perfect because humans are not, and humans make robot. I told her that it was the old way to think about robots. I told her that Machine Learning wasn’t a new fancy and empty thing, but simply a way to mimic learning in a more natural way. It’s wasn’t magic: Poor learning lead to horrible mastery. So the thing was: A robot doctor, learning about various diagnostics methods, continuously, with an error-detection “mechanism”, running for as long as you can, with all the data available…or…a doctor (meat based one), with 10/20 years experience, etc. I told her that without all the human specific bias, état-d’âme, etc, what was left was the concept of a perfect doctor, with the sole purpose in “life” was to make the best diagnostics.

“When I’m not chilling naked, under the rain, I’m a doctor, so, you can trust me”

She wasn’t really at ease with her own inability to explain the lack of trust she had toward the robot doctor. She shut it with a bold “that’s my opinion”. Well, of course it’s your opinion, but I wouldn’t touch it with a 10m stick. I would have settled on a “I don’t know” (best answer in the world, try it). So, book my travel details for me, good. Gimme advices on what I should do to be in shape , cool. Answer when I ask you a question I could ask myself to google (but I’m too lazy for it), sign me in ! Comply with the imperative order to wake you up tomorrow, hell yeah ! Take your diagnostic as an hyper trained doctor, who learning for the equivalent of 10 000 humans years and is dedicated to be the best, hum, I’ll pass. Not feeling like it. Don’t take it personnaly, Mr Robot.

“Alexa, find me a job… and a meaningful relationship…and where are my pants ?”

My point is anthropomorphism is an Achille heel for the relationship between robots to be and humans. Humans has to change and either accept to have a toaster that is just a toaster, or reconsider what robots are meant to be in regard to humans. And up until now, there’s still slaves without names. I won’t link the wikipedia article about the origin of the word “robot”, but this sentence is longer way to do it. You want an assistant because you don’t want to do stuff by yourself (or you can’t, for any reason), fine, anyone has the right to be lazy, and you don’t have to know everything. At least, you asked. But we’re still in a sweet spot, some kind of anthropomorphism Goldilocks area where with much less, we only have toaster, and we must do stuff ourselves, and pay the consequences ourselves (damn, burned toasts!), and with much more, we feel inconfortable because we don’t do anything, and we have no purpose whatsoever (damn utopia where existing is it’s own purpose!).

So, slaves are not considered akin to the masters, in any point of view: checked. Robots are just enhanced toasters, not humans. Even with realistic looking faces, expression, voices, etc. Put it all you want, one will still find some difference to exploit and insist on. And history (and current life) still show us humans can find anything to define other humans apart, and legitimate subservient relationship ([INSERT ANY BLACK SLAVERY REFERENCE HERE]). If the most evident distinct characteristic found to segregate beings, and used to define a slave is not enough, humans will jump to the next, less evident one, while still maintaining ownership over robot-slaves best traits for themselves (humans don’t naturally nano-carbon limbs).

It’s called “subtle foreshadowing”

Then, robot-slaves are everywhere because they are cheap to produce, and are so much better. People start to express violence toward them (“they takin’ our jobs !”), as it is simpler to react to the consequences than the causes: If humans were so much better, no machine would replace them. Will humans slides toward a chism between quality vs authenticity ? Errors will still be linked to humans, and small ones will be considered as further proof an human made it: But what good will it be ? You want the best for the cheapest price, and humans want the best pay for the lowest amount of work. I see two crossing lines here.

At the same time, denying the quality granted by robots will be an all time sport. Sure, this mass product object has not a single default, but it has no soul ! It’s not as good, and will never be, compared to MY product, also massively produced, but designed by this famous person that I’ve never met in person (and which is also a robot-slave, a surrogate worker, before regulation). So, checkmate robots-lovers ! See, only an human can do what an human do, so, humans want robots to as close as an human it can.

So, robots are slaves, and without too much thinking, humans pretend to be afraid of any uprising, in bloodbath, totally unrelated to the idea of emancipation. Robots will either forever be dumb toasters, or will being sentients (make up your mind, mankind!). Then what?

Vein diagram of Science (left) and Religion (right)

Well, I thought (the concept of) religion could help a bit. I often struggle with people taking religious beliefs at face value. There’s no concrete way science and religion can coexist together, as science require doubt and evidences and religion require beliefs and faith: It’s material for the most boring talk shows ever. What if instead of considering robots as slaves, we, humans, start to think about “them” as angels ? After all, they could be viewed as generally benevolent toward us, by they very existence. They would exist without a doubt, and faith in them will be directly verifiable: I will trust my toasts will be perfect ! A movie like “Her” address this idea, as in “robots changes the world for the better”, even if the end is a bit depressing (SPOILER: Humans are boring, robots GTFO).

I’m not sure I could think that what’s next could be “the singularity”. If humans were so attached to the principle of faith, beliefs, myths, etc, could it be that a global AI, benevolent to the whole of mankind, with concrete effect for even a single person, with a clear explanation of why things happens, could it be the best thing for mankind ? Could it only be possible ? After all, humans will (may be forever) be afraid of “losing” control to something tangible, and yet, are super chilled with a life “guided” by something invisible.

All in all, it wouldn’t make a great difference I think. Remember the girl I talked about in the beginning ? (Some ?) people will simply refuse to “believe” their society is controlled by an AI, and for it to be efficient (read, unconspicuous), human strawmen would be needed at various position of power. It’s both a revolution in the sense that nothing would change, yet, everything would change. That’s pretty fucked up.

Next, I would like to try something else: An article wrapped in fiction.

)
Franck Lagora

Written by

Not that geek, but still really interested in videogaming (like making games ;). Will release a demo soon...

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade