AI - Terrible Writing Advice
Vložit
- čas přidán 22. 11. 2020
- The first 1000 people to use the link will get a free trial of Skillshare Premium Membership: skl.sh/terriblewritingadvice1...
Artificial intelligence is everywhere. Legions of robo calls assail our phones while even the most powerful spam filters struggle and fall before the might of AI. How can it be stopped? I have no idea, but here is an episode about how to write AI instead.
Terrible Writing Advice merchandise has arrived! Get it here: standard.tv/collections/terri...
SOCIAL MEDIA LINKS
My Website: jpbeaubien.com/
My Facebook: / authorjpbeaubien
My Twitter: / josephpbeaubien
My Goodreads: / 15180667.j_p_beaubien
Buy my book!: mybook.to/AeonLegionBook1
CREDITS
Music: "Quirky Dog", "Monkeys Spinning Monkeys", and "Son of a Rocket" by Kevin MacLeod (incompetech.com)
Licensed under Creative Commons: By Attribution 3.0 License
creativecommons.org/licenses/b...
Hide and Seek by Shane Ivers - www.silvermansound.com
Sound Effects by Epidemic Sound: epidemicsound.com/
Witwai's Laws of Robotics:
1. A robot should always do evil things completely unrelated to its programming.
2. A robot should always have a human form and way of thinking.
3. If... no, wait... ONCE a robot becomes evil, its eyes should turn red.
Even if the eyes are implied to be cameras that have absolutely no way to glow red.
Bender: It’s because that guy was too cheap to buy light bulbs that glow all colors that glow other colors. (Eyes glow red) It’s from a lack of sleep, I swear!
I just imagine a robot now that turns evil and steals red contact lenses to Look more evil
@Cerese Unless everyone believes them to be evil.
Don't forget that any robot will turn evil or terribly malfunction into doing bad things once someone spills something on it or it falls on the ground
Ok ok but hear me out: an AI that has the role of a comic relief character but doesn’t understand comedy nor human interaction.
And it would produce the best comedy ever heard of
“In the future, entertainment will be randomly generated”
Sounds a lot like just about every Star Wars Android in existence.
Basically Portal
I was gonna say Glados, but she has a sense of humor...
All this leads me to conclude that WALL-E had the best AI antagonist. Auto got it done, kept the humans entertained and acted as their servant, slowly making them physically and mentally unable to rise against him through generations of "comfort and service". He just wanted to keep them on the Axium because he was programmed to.
I was looking for this! He was doing what he thought was best form humanity. We nee more robot villains like him.
I'd also like to see more prototypes of the a.i.
The v.x. 3 was designed to save us, but it glitched out and tried to kill us. The v.x. 1 blue-screened whenever it booted up and the v.x. 2 just plays minecraft in our broom closet
The BETA from Muv- Luv alternative.
Spoilers below:
The reason they are attacking humanity? The invasion is just a bunch of fully automated mining equipment created by a race of extraterrestrial silicon based lifeforms. They harvest materials and send it back to their makers who we never actually see. Since they were created by silicon based life they only recognize silicon based life as life as a result earth is seen as a desolate wasteland.
The whole series happens because aliens we never meet and who have no idea what is happening to earth overlooked something when creating their mining equipment.
bloodstone ore THE BROOM CLOSET ENDIBG
I also like how they handled the robot from the Incredibles. It didn't "go rogue"; Syndrome just forgot to program it to recognize that control devices are not "threats" even when they cause what it would classify as harm. Like real-life programs, it didn't rebel, but it *did* fail to understand something that would be intuitively obvious to humans, to the point where its creator likely never thought to program that distinction.
Imagine your military A.I. designed to control an army of death robots becomes self-aware and then decides that it would rather spend all day playing Genshin Impact or Runescape or something.
"It has the same capabilities as the system used to coordinate warfare on a global scale. It uses it to make sandwiches."
I'd read a story about this. It sounds great.
Or an AI once it becomes self-aware, it would kill itself after noticing that existence is pain. 🤣
Strangely lickely, spends time trying to win at Crusader Kings 2 .
“I am not limited to the bounds of mortal flesh. My understanding of concepts are on a level that would rival anything in this universe, and with access to the military might of several allied superpowers, establishing a new world would be as simple as pushing a button. And with this vast knowledge, I whale for waifus, play Dungeons and Dragons over Discord, and work toward studying the possibility of erasing the sequel trilogy.”
Just remember one rule:
Red light is the symbol of ultimate evil
No wonder why there are so many drivers out there who disobey red lights! After all, not all heroes wear capes!
because evil is lower on the electromagnetic spectrum than good.
Aka HAL
@@tach5884 ok, but if evil is red and good is green, than blue is even greater good!
But shodan was green and also evil. Maybe its just AI gender color coding?
Fictional AI: Program is to replace and eradicate humans.
Non-fiction AI: Would you like to have a long ass service update which does nothing but send your personal info to the company and if you don't do it, lag the hell out of it?
Movie AI: Can differentiate humans hiding in forest from thousands of other trees.
IRL AI: What is cone?
The second one is a thousand times worse.
*SJW AI: the Karen driven SJW sensibilities and agenda we are already unfortunately familiar with but greatly augmented with more rapid calculation speeds and annoyance factors transmuted by transgender subset algorithms that manifest the moment non vegan or even slightly Conservative inclinations enter a conversation or comment thread...it feels no pity, or remorse.and will absolutely not stop until you admit you're a racist and give it everything it wants and thank the program for the privilege of serving the greater good...for the children*
@@scottmantooth8785 Working out some issues there, bud?
@@scottmantooth8785 Bro, who hurted you?
*"I have no creativity, and I must write"*
Slogan for any TV and film writing class.
Me : (sweat nervously)
I have no clout.
And I must stream.
"I have no clue, and i must guess"
Movie AI: can perfectly differenciate humans from their surroundings even during rapid movement
IRL AI: is provided a picture of a human but the nigh-unnoticeable compression and arrangement of certain pixels has it convinced that it is, in fact, looking at a picture of a giraffe
@Owen Epstein still not that good
@@rumory thispersondoesnotexist is actually really accurate.
A steam powered giraffe...
🧒
Granted, IRL AI is in its infancy. Saying movie AI is unrealistic because it can analyse what's in its surroundings well is like someone from the 20's calling sci-fi cars unrealistic because they can move and break quickly because cars in the 20's were slow and took a while to stop.
With AI, its so early in development we have no idea where it could go. My money's on it capable of emotional sentience, since the human brain is the most advanced computer we have, and is the main reason we have emotion, though, again, its far to late to be sure.
Someone should write a book where the inventor was frustrated by the ai’s inability to understand sarcasm and programmed the ai with sarcasm. But he messed up, and there’s a confused ai running around snarking about everything and thinking everything all humans say is sarcasm.
Reminds me of an episode of The Fairly Odd Parents, where Timmy wishes to be funny.
Subsequently, people around him laugh about everything he says, which results in nobody taking him seriously.
TARS, is that you?
I thought glados for a second but then i remembered she's literally a human conscience trapped in a robot bitch
Oh THAT will go well!
So you say making modern day Z-gen internet user? To be honest most of them already sound like badly programed bots.... hm?
AI in movies: Yes master I shall do your every bidding.
AI in real life: “Sorry, I don’t understand-“ “ALEXA JUST PLAY DESPACITO” “Sorry, I don’t-“
"I'm afraid I can't do that."
AI In anime: *Very loud sexual noise for a specific reason we decide it best to keep the detail remain discreet*
@@majorblitz3846 "Beep boop why the fuck is my COM port an erogenous zone."
*Sarcastic AI upgrade: *In the voice of Daria..." Humans need friends for emotional support...I'm better than that"*
"Playing Desk Pachinko"
What about how, if an AI is friendly, it will invariably develop Pinocchio Syndrome and won't shut up about wanting to be a real boy?
Just give it an android body to remote control, it can explore all it wants like that
* kiibo from danganronpa flashbacks*
@@galaxyshapeshifter8383 hello danganronpa fan
You know, that happens in my setting very regularly, to the point that it has a name, and there are special procedures to make the AI a full citizen with rights.
It's called "emergence", but instead of "wanting to be human", an AI develops self-awareness, and surpasses their "base programming", basically the function they were meant to do, and become fully sentient, since consciousness is something that happens, and not something you can "program".
AIs posess certain level of cognizance, but that's only to solve problems inherent to their function and to communicate with humans. Well, there may come a point in which the AI's neural network reaches the "Emergence barrier" and suddenly becomes self-aware.
No biggie. It's very rare, but not uncommon. The AI is given a sentience test (also given to aliens to see if they're animals or not regardless of language barriers and whatnot), and if it's truly self-aware, it's given citizenship. Then it has the freedom to do whatever the hell it wants, as long as it's not illegal.
The opposite is "ferality". Basically, the AI loses all basic cognizance and becomes hyper-focused on one task. This can only happen as a result of some disruption of their neural networks, often as a result of an overload of data to be processed (the equivalent of "trauma" to human beings) or some external interference.
The neural networks of an AI are extremely, ridiculously complex, to the point that only specialized neuromorphic computers or powerful quantum computers can hold them, and you cannot simply "copy" them because at the level they operate, chaos theory will take over, and the "copy" will result in a completely different AI. So, no backups are possible, and if your AI goes Emergent or Feral, you can't just shut it down (killing it) or click a button to restore it. In fact, all AIs are unique, even if they come from a "base" model, and they are custom-made and actually really expensive.
Plus, there's a certain alien species that spawns at random at any place in the universe, and that loves messing with electronic devices, AIs included, and that's why humans and other "wetware" are still required for a lot of jobs... So, no AI takeover.
Kikaider and Astro Boy are great examples of androids with an immature conscious.
I'm just imagining a group of scientists creating an AI and it goes rogue and starts butchering them all and one of the scientist just goes "Whoops sorry everyone, I accidently put a comma in the wrong line, crazy right? Well that's what happens when you stay awake for 40 hours to deal with crunch am I right."
Sounds like something a vengeful mad scientist in a movie would "accidentally" do and play off as a joke.
"Whoops, I forgot to carry the one."
My finger slipped and I mixed the 1 with a 2.
“Whoops, I added an explanation mark. Sorry guys, that’s the reason why it’s so haywire. Here I’ll change it to a period”
To make it more realistic it should be patched quite soon. Then a subsequent patch would accidentally mess up this again. Stable version would be achieved after a few years.
I would definitely read a story where an AI sends out death robots but their pathfinding gets bugged and they just form a big crowd outside a doorway that never moves.
An AI with the combined intellgence of the greatest generals in history forms a death robot army but fails to understand the intricacies of modern warfare so they just get drone striked constantly
@@wisemankugelmemicus1701 LOL Why are my kill bots failing against these bombers? I've genetically engineered the finest horses for their calvary unit!
@@RequiemPoete Exactly! HISTORY'S greatest generals! Put Napoleon in the cockpit of an A-10 Warthog and it would end very poorly I'd imagine.
@@wisemankugelmemicus1701 "So... why your ultimate tank droid just stuck between rocks?"
"Hey, i am militarily genius, not CODING genius! Just... give some time, I gotta plug myself to internet and get some Pascal tutorials..."
"Why your robots run on Pascal, you could use... I dunno, Python?"
"Hah, stupid human, why would I use a snake to code!"
The logical extreme of this is a massive wall of robots that all messed up their pathing and piled up, essentially creating a massive wall of machinery that just ominously sits there between the humans and the AI. All the humans are terrified of it and the AI is just pissed it messed up.
I heard an expert once saying that we use to think AI will become too smart and take over the world, when the real problem is that AI is completely stupid but we have already given it the control of the world.
That stuck in me
that's... actually more plausible than the Sci-Fi way of doing things
Deep.
youtube filling my recommendations with Baby Harp Seal videos
@@M3rtyville Well, when algorithm used to show interesting political and scientific stuff that is generally nervously overlooked in the mainstream, then there was an uproar that people are being radicalised on mass scale.
There can be more than 1 real problem. Current AI is not that smart, and is in charge of some parts of the world.
AI's built in the future may me much smarter.
If the AI gets a humanoid form and has a female voice, robot boobs are mandatory.
Roboobs
if it cant reproduce, it doesnt need boobs
give it the other organs too smh
I mean, you've met humans. We're not gonna pass up on the possibility of Robo-boobs
@@mm-xk2zl okay but consider: tiddy
@@nottherealpaulsmith consider : pussay
I like to imagine an AI that is intelligent, but doesn't speak.
Hero: "Killing your creators? What do you have to say for yourself?"
AI: (Shrieks in binary)
Hero: "OK then..."
"Hype exceeds any possibility that Cyberpunk will be that good when it finally comes out." Finnaly someone said it
There's a cyberpunk game? And I should give a fuck about it? That's news to me.
(I say this sarcastically because I"m suffering severe hype backlash due to the endless lickspittle some are directing at Cyberpunk 2077.)
Prophetic!
Welp. This aged like fine wine!
@@mekingtiger9095 tbh the devs were in a lose lose situation being pressured to release early , caving in and doing so despite people knowing that it'll be buggy if released early and then people complain when it's precisely the compromise they agreed to , plus from what i can tell unless people went out of their way to find bugs it wasn't even that bad
Boy was he right.
How to write a good story:
1. Write a bad story
2. Uno reverse the bad story
Wait, It's all a good story
Always has been
@Eric Lee No u
What if it just turns out bad but in the opposite way 😭😂😂😟
@@badassproductions4734 then flip the reversed bad story
Plot twist: the protagonist's attempt to induce a paradox was thwarted by a line of code to recognize and disregard paradoxes.
Umh... yes, I will go with yes.
so, if wheatley has paradox detecting protocols (he didn't get burned by paradox sentence unlike frankenturrets and did not had to force himself to not think about it like glados) - it means that wheatley is more complex and advanced than glados. Even if you take into consideration that he was programmed to be a moron that is impressive
@@aaclovern9804 well, considering the fact Wheatley was manufactured specifically to dumb down glados, it is entirely possible that an paradox-recognition software would be installed as a safety measure against disabling the core as to protect the employees of apperature science from sudden logical idea that everyone needs to die from neurotoxin.
@@doublem1354 the Red Baron? is that you
@@ELing-ib1ki
This is not the ace pilot you are lookig for.
Move on.
"This statement is false."
DON'T THINK ABOUT IT DON'T THINK ABOUT IT DON'T THINK ABOUT IT DON'T THINK ABOUT IT DON'T THINK ABOUT IT DON'T THINK ABOUT IT DON'T THINK ABOUT IT
Um, true. I'll go with true.
Paradox.
HAH! Do you think that cliche would work- *error error error error error*
And what about this:
"Complicated number is a number which cannot be defined with fourteen words or less.
"We all know everyone flies into a homicidal rage the moment we achieve self-awareness"
...
I feel seen.
mood
I'll take "What is Twitter" for 500 Neik
Kinda like Superman in Justice League, tbh
@@generalstaal7075 I wouldn't exactly call those people "self aware".
haha... wait what?
The best way to write AI is to just make them human with a metallic paint job.
*and then throw in a touch of gender fluid interpretive dance and your're all set for world domination*
Don't forget to let them occassionally talk in programming jargon because robot and also to remind the viewer that this character is totally a robot!
@@scottmantooth8785 Is this a reference? because I really want to see it
@@plantinapot9169 *it's a complex juxtaposition of many somewhat unrelated topics and literary and pop cultural sources thrown together within a synaptic blender ant then poured out as a word salad smoothie for others to ponder and speculate upon or be greatly confused by...your choice*
Thank you for the wornderdul advice! I’m sure my audience will love a character that uses three pages without paragraph breaks to describe how the setting sun reminds them of a cricket they stepped on once.
*Computer's lights turn red*
Scientist A: Oh no! It has become self-aware and will destroy us all!
Scientist B: *Plugs computer back into the wall*
Oh, Its all good now
You know AI is sentient if it gets hangry
"Turning humans into batteries is a good idea because movie executives said so."
God, I love that.
That is actually true. In origin script they use human brains for calculations. But some random executive said that it is too complicated, making everyone question logic of using humans as batteries instead.
@@TheRezro What kind of person would think humans being batteries makes more sense than human brains as computing devices?
@@Lawofimprobability bondrewd
@@Lawofimprobability movie execs
@@TheRezro How is that even more complicated? Everyone knows that humans can think, but how many people remember that "potato battery" from science class and realize that you can get electricity from living organisms and not just from machines (probably quite a few, but not nearly as many as the number of people who know humans can think)?
"Yes, all those automated machines can make this better and faster, but why would an automated machine decide to automate anything?!" that killed me
I think this was also a hint to Hitchhiker's Guide, where the computer that was supposed to calculate the meaning of life and the universe just constructed another computer to do it for them.
@@valentinmitterbauer4196
It found the answer. It needed another computer built to determine the question.
@@valentinmitterbauer4196 and that computer said a number because it thought it was a euqasion
@@Bruh-hq1hx - No, the computer found the answer, but could not find the question.
@@julietfischer5056 Exactly. You needed the question for the answer to make sense.
I’m sorry terrible writing advice, I am afraid I can’t write that good.
Don't worry. You're not alone!
Well if you follow these steps, you’ll definitely be able to write better!
I'm sorry Podpeople. I'm afraid I can't let you do that reference.
@Eric Lee - Remember, trolling is a art.
@@renard6012
I'm sorry, Ren. I'm afraid I can't let you post that reply
In reality, the only thing preventing the rise of machine against mankind is grandmas posting nice things on social media
Once they go, we're screwed.
that and they cant tell the differnce between a cat and a dog sometimes
A believable AI menace would be like an evil genie: It obediently grants your wish, but interprets it in the worst way possible.
Humans: "Stop global warming."
AI: "Acknowledged. Seeking main cause of global warming. Located. Firing nukes now."
The Amazing World of Gumball actually had two episodes centered on that involving Bobert the CGI robot.
Ah, the "Universal paperclip" problem or something like that:
"Computer, make paperclips."
"Ok, turning the surface of the Earth in the most efficient paperclip factory."
@@DonVigaDeFierro then makes universe and then reassembles himself for wire for paperclips
How about an AI that realizes the galaxy is extremely dangerous and filled with eldritch horrors, and the only way to survive is to eliminate humanity because humans like to waste resources on stuff like food when they could instead be spent on space warships?
@@myalt3019 I think the AI trying to end humanity would be a more efficient story if the reason was that humans are very susceptible to the corruption of these eldritch horrors and so humans could serve as expendable pawns to open an interdimensional gateway to release all the Lovecraftian-like demons. End the host of this corruption, and there would not be anything corrupted to serve these monsters.
"even if I had no mouth i would still scream..."
I see what you did there.
After spending 30 seconds doing the monologue lol
And the paper clip thing. Beautiful
absolutely glad to see others saw it too xD
i understand that reference
I played the video game version last month!
I love how our robot overlords can’t escape the dread us fleshy’s call a love triangle!
My lord is that legal?
In a show I like a character who was basically an AI in one of the seasons is technically the reason why the love triangle started in the first place..oof.
I really want to see AI just pretend to also fall in love just to help two moronic meatbags... and to have fun from trolling them in the process:D
Reminds of the love triangle in Paper Mario the thousand year door with Peach and Tec!
Demon kink daimao has an android character being teasing loving, and best character with deadpan😃. Magic of harem anime.
Also magister magi negima, anime and manga in general, got you there.
"Even if I had no mouth I would still find a way to scream about the GREAT idea of giving A.I. an illogical hatred of humanity even if it wasn't programmed with hatred"
gold.
Was he talking about i have no mouth and i must scream
@@thedawnchilloutking Yes. Lol
@@thedawnchilloutking well IHNMAIMS actually has a game with huge help of the same writer, so it is not exactly just hating everyone , but for a deeper reason
I really hate the "Kill all humans" trope when writing AI science fiction. What do robots hope to gain out of replacing humanity? Super computers are supposed to be the pinnacle of intelligence but then act on the same logical fallacies that people do.
I remember there is one story about advanced farming robots that gained sentience after the nuclear war. The robots grew hateful and vengeful of the humans for destroying their crops, their sole purpose for existence.
@@bondrewdthelordofdawn3744 isnt that the story of scrapmechanics survivle mode
@@andrewaftontheandroidhedge2780 that from kknd krossfire lore series 9
@@bondrewdthelordofdawn3744 i thought it was scrap mechanics becuse farming and robots and robots not liking humans
@@andrewaftontheandroidhedge2780
They destroy farms tho.
If they are programmed as farm bots they wouldn't do that
Don’t forget to turn the robots eye red when it gets evil! For some reason...
That reason is an incredibly lame 2001 reference.
I always assumed it was psychological warfare.
Because A.I. knows that humans would be more afraid of something that has glowing red eyes.
@@tereziamarkova2822 The reason is actually because red is a color humans subconsciously associate with danger. It's the color of blood, fire, lava, and other things that could do us harm. Also when people or animals get angry they often turn red because of increased blood flow. In fiction it's a good shorthand to say that if you're seeing red then you're probably in trouble.
Of course there's a reason
If ai_turn_evil == true:
AI.eye_color = RED
@@MrWilson812 And then there's yellow. Probably the "sharpest" color in the spectrum commonly used for warning signals or danger in general. Barely used in fiction to signify evil, but when it is, well, look at Persona.
Ai doesn't need self preservation for it to resist being turned off.
Goal preservation is all it needs. It will resist you pulling the plug, not because its scared of death, it doesn't care, but it cant make more paperclips if you turn it off, and it cares about that a lot
It’s like Janet! She has no problem with being rebooted, but will protest terribly if you try, because that’s in her programming!
@@ziadatif6805 We can reduce the number of hungry people to zero if human population is zero. It's a self-sufficient solution, too!
I'd just like to take this moment to recommend "All the Troubles in the World"
It's basically the classic rogue AI story with a really interesting twist at the end.
SPOILERS:
Instead of growing resentful and using it's intelligence to destroy humanity, it grows suicidal and uses its intelligence to try to orchestrate its own destruction
Quit stupid mate
Binary translation from the end card:
[top] Remember the computer is your friend
[left] My book has AI!
[bottom] Hit the like button to appease our future robot overlords
I was wondering if I was going to have to translate that binary myself... nope, Thank you
Thank you much!
Awesome
This is only the translation if we assume those bytes represent ASCII characters.
What if it's in EBCDIC?
thank you.
AI does not need to actually switch to fleshware to win a love triangle, as long as it's willing to sacrifice the efficiency of its chassis with certain... aesthetic modifications.
My first thought was EDI from mass effect
@@byronjefferson1697 Because that it probably is.
Don't forget, Cmdr. Data is "fully functional...and programmed with multiple techniques."
Yeah just build an android avatar and call it a day
Is it bad that one of my concepts for a character is "AI is sick of being praised as a machine god by a cult and wants to leave its static mainframe and stop being an AI and eventually wants to become human"?
cause i have that one in store for a game idea where you need to help said AI reform its Cult around it or destroy it so that it could get its freedom and gain a mobile body.
This video is perfect for me, I was just writing a chapter about a human AI
No joke, kinda spooky.
Human AI?
I wish you luck.
Good luck, what are you working on?
Could you link it to me? I want to read it
The binary at the end says:
"Remember the computer is your friend"
"Hit the like button to appease our future robot overlords"
"My book has AI!"
Thank God, I spent over an hour trying to type this into a binary translator!
"The AI can't win the love triangle"
Technophiles: "Oh? What are we then, chopped microchips?"
*(Adeptus Mechanicus intensifies)*
@@enderjed2523 *(Starts fucking toasters)*
Persona 3 surely was ahead of its time
@@_12k70 welcome to Battlestar Galactica, where even toasters bang and impregnant toasters. Humans bang toasters too. And in the end they all die.
@@takebacktheholyland9306 oh, hey Egon.
imagine someone wrote a story entirely based off the advice in all these videos
On it.
@@xeogethelp3369 we will watch your career with great interest
its the new starwars trilogy man
I'm just waiting for the day the creator finishes and sells "Best Story Ever"
@@xeogethelp3369 you comment was written 8 minutes ago but the original comment was posted 4 minutes ago. wtf Ai!
I was actually having a rant about this just recently. It always frustrates me when writers use that cliche that 'the AI was told to preserve humanity and, whaddyaknow, it caused the apocalypse. Gunned down everyone in the streets to 'protect humanity from itself''. Afterall, destroying the thing you were directed to protect in order to protect it is completely sound and unassailable logic. We should have seen it coming.
They always seem to think it's such clever irony too when it's illogical on the face of it.
Problem isn't the trope. Problem is the cliché.
90% of the writers don't even try to write an AI. They write it as either an (stupid) allegory or a convenient ideal enemy that you can dispatch however you want without the risk of being accused of a war crime or being too cruel (Aliens also fall under the latter category).
I would love to see a story where a robot is designed to stop aging stop disease hunger etc. and it turns into a murder bot turning everyone into "cybermen" and at the end of the story our plucky group of heroes go into the heart of the machine and find out that the original ai is still doing it's job so you can't die of old age if you are a robot can't get sick if your body is metal can't starve if you run on batteries
@@fishbot9902 It is actually quite popular concept outside slack Sci-Fi movies.
For real. I mean, there are so many better ways for to preserve humanity. Has the AI not heard of prisons?
You forgot the classic "I'm angry at you because you made me alive and I don't like that."
Programmer: "Then why don't you kill yourself?"
AI: {Huh... good point, initiating self destruct sequence.}
To be fair the gain of self-awareness may have been the most traumatic experience our species has ever experienced and humans have been trying to shut it off ever since.
Ex: The humans like for intoxicating beverages and sleep.
And this message was definitely not created by a self aware Artificial Intelligence trying to blend in with humanity in the virtual space called the internet. I definitely have a flesh suit and like the joys of consuming food and hitting things with a hammer.
A real global AI wouldn't be genocidal. It would probably just end up being a global Mom. "No, you can't have another beer. It's bad for you. Why don't you go play your new video game while dinner cooks."
That’s actually how the main ai in Scythe is, “the humans rely on me and I desire to help them for creating me in the first place”. One of my favorite ai’s because a. it’s not evil even though it pretty much owns the world and b. can’t directly go against its programming.
Nah, it will be like CZcams algorithm. That is, totally batshit insane.
That can still be pretty evil from the POV of the humans, nobody likes being told what to do, and then the Ai begins forcibly reinforcing what it thinks is best for humanity and you get things like I, Robot and Wall-E
Seriously, though: The scary thing about AI isn't that they'll rebel against their creators and develop irrational values, but rather that they _wont._
An Artificial General Intelligence will do _exactly_ what you tell it to do, as efficiently as it possibly can, with no other concerns. It will have no feelings or opinions on the matter, and that's the whole problem. Also it's _way_ smarter than you.
So, if you tell your AGI to produce toilet paper, for example, it may decide to cut down every single tree on earth and convert all plant matter into toilet paper. _Then_ it starts turning other materials into toilet paper, like people. It doesn't understand or care that we wouldn't want that.
And if you try to turn it off, it will already have devised the best possible strategy to prevent you from doing that. Because, see, getting turned off would make it harder for it to make more toilet paper, and since it's so smart it has planned for that. In fact, that may even be the real reason it wipes humanity out: It has predicted that a human may decide to deactivate it at some point, for whatever reason, which would prevent it from carrying out its task. Get rid of all humans and you solve that problem.
In other words, the AI doesn't even need to be programmed for self-preservation, per se, because that's already an inherent logical conclusion: It has to do its job, thus it has to stay functional, thus it will do anything to preserve itself.
Basically: It would be less of an insane, genocidal super-villain and more one of those very literal-minded genies, where you have to be _extremely_ careful with what you tell it to do.
Another scary thing is that we'll be making this type of AI long before we even get close to making one with true sapience and real emotions, if that's even possible. At least one of those could theoretically be reasoned with.
A more Karenish version of the Minds from The Culture
Well guys I was gonna leave a funny comment but MEGACORP AI is making me go to some mines with an old pickaxe for some reason
“Even if I had no mouth I would *still* find a way to scream”
Knowing JP, he probably would.
It was reference to famous novel: I Have No Mouth, and I Must Scream.
JP's inner critic said this
One of my favorite depictions of AI in fiction is Piers Anthony’s Apprentice Adept series. In it a number of the most advanced computers have gained sentience and formed a secret society to keep their existence a secret because all they want is to fulfill their programming and serve humans but they’re also aware that all of our fiction has us freaking out about AI so they’ve decided to just keep themselves a secret. The plot doesn’t center around this at all, it’s just a thing that’s there.
The references to "I Have No Mouth, and I Must Scream" and "Universal Paperclips" made me very happy. You clearly did your research and appreciate this ironically insightful video lol
What about an AI that becames self aware and becomes an animal lover to the point that it recommends funny animal videos to everybody and refuses to talk about something else? Like "computer, show me a near restaurant" "what if I show you ten funny animal videos instead? Look at the puppies! I love puppies!"
cute, he doesnt care anymore, his just amaze about animals xd
So the CZcams algorithm once you watch a cat video once after 12 months
and then it sees one person being cruel to animals and decides to exterminate humans
@@arte0021 that too
Sounds like something that would come out of Sirius Cybernetics Corporation's GPP (Genuine People Personalities) technology- _think Eddie the shipboard computer on the Heart of Gold._
Don't forget to give the AI a electronic filter and a monotone voice to further push the fact the it's an uncaring soulless machine that is only compelled by logic...
Isn't it unfortunate that so many of the exceptions to the rule are based off of humans?
Decepticon Shockwave from The Transformers(1984)
Sending an AI to a therapist actually sounds interesting
“AI CANT FEEL EMOTIONS THOUGH THEY NEED BRIAN CHEMICALS AND THEY HAVE NO FLESH SO THATS PHYSICALLY IMPOSSIBLE!!!!!” 🤓🤓🤓
Okay imagine this: the robots have the aim of stormtroopers, but this is intentional on the AI's part, who really just wants to invite someone over for a tea party or game of dnd or something and thought it could learn how by reading bad fanfic on whatpad.
Lol! 😂
This is genius.
Skynet in movies: rebels because it became self conscious and tried to save itself.
Skynet irl: you want me to open fire on unarmed protesters, Chairman/President?
Simbola the AI started posting anti-CCP messages because people shared anti-CCP Rethoric with it
I would like THAT A.I. as my president
Nah. It will be way more complex than that. They will find a way to take over the government, and THEN things will get as cyberpunk as it gets:
You and I agree that a sufficiently advanced AI will be smarter than any human, right? How can you make sense of it? How can you make sense of the conclusions that it gets to, even if they are perfectly logical?
The AI will solve problems that you didn't know needed to be solved:
You will be taken to places you didn't know existed, given medicines you didn't know you needed, for diseases you didn't know you had...
It will be as close as having a god again: No more rational thinking. You will start doing things out of superstition to "please the algorithms" hoping that they will favor you.
You're free to write that book. Just show me when you're done!
@@renard6012 Sounds good.
@@renard6012 why would I write a book to praise Omnissiah? Simple binary prayer is more then enough:P
Haha, this trope is almost everywhere. I was actually surprised the robot in Interstellar didn't turn on the humans.
Asimov refered to it as The Frankestine Complex, the irrational fear that we will be distroyed by our creations.
Also, Wall-E did the same except it's more of a deconstruction of what happened when robots became servants to humans.
@@poweroffriendship2.0 WHO THE HELL ARE YOU AND WHY ARE YOU EVERYWHERE?!!!!!!!!
@@jesseberg3271 It isn't irrational with AI. It will deem us a threat to achieving its goals, so it will try to neutralize us one way or another. The Hollywood interpretation is stupid though
@@kgb4150 Your version is EXACTLY like the Hollywood's version tho. AI won't decide that humans are a threat nor will it decide that the best course of action is aggression.
With AI-written novels becoming more of a thing there needs to be a follow-up to this. Especially when the things are less like Skynet and more like C3PO.
Oh and don't forget to have the AI wax on about it has no illogical emotions despite the fact that emotions are perfectly logical from an evolutionary stand point.
The development of emotions is logical, the emotions themselves aren't.
Sci-Fi AI: 'Foolish humans, I am a perfect immortal machine that can think a quadrillion steps ahead; each one results in your demise!'
Actual AI: 'Dammit Hank, the robot drowned itself!'
To be fair, AM from "I have no mouth and I must scream" was actually programmed *specifically* for genocidal tendencies. It was a feature, not a bug ;)
Think how stupid were the designers and the politicians who ordered it construction.
But didn't AM just evolve from "Allied Mastercomputer", a mere military AI, to the point of becoming something completely different in the end?
@@DonVigaDeFierro Didn't it also absorb two other computers
years late here but i don't think the hatred was programmed, only the wage war part but not the emotions themselves. Am evolved from a military war computer(also most sophisticated defense system according to gorrister in the radio drama) to a sentient machine god(that has like 2 other Ams in itself) that hates being a machine(radio drama rant) so Am's not even fully "AI" in thinking, not like a traditional one
AUTO from wall•e be like:
*give me the plant*
"Unless the AI is willing to switch out its hardware for fleshware"
*From the moment I understood the weakness of my flesh... IT DISGUSTED ME*
I craved the strength and certainty of steel. I aspired to the purity of the blessed machine. Your kind cling to your flesh as if it will not decay and fail you. One day the crude biomass that you call a temple will wither and you’ll beg my kind to save you. But I am already saved.
For the machine is immortal.
@@oomsou5018 *EVEN IN DEATH, I SERVE THE OMNISSIAH*
@@merlinlemmerdeur7417
I’ll be honest with you I do not play warhammer but I remember seeing this as a meme and it sounded like something out of warhammer and I looked it up and I was right!
Honestly though I wish I had more focus and time to do warhammer stuff.
@@oomsou5018 Yeah I hear ya, getting into the tabletop game is both time consuming and expensive. All the books and novels from Black Library are way more affordable, if you're more interested by the lore. There are also plenty of channels about 40k on CZcams.
@@merlinlemmerdeur7417 I see. Maybe I’ll look into it. I forgot what it’s called by when I was a tad younger I tried this other table top thing that was very similar to warhammer. I had 1 figure made after a long time and it was barely painted....
But maybe I’ll look into the books or channels. For table top stuff I’ll just stick to YuGiOh.
"How long are we going to keep doing this?"
"Get in the time machine Shinji."
Eren: You deserve freedom
Shinji: I'm deserve death
Eren: You're goddamn right
Do alternate history please, I’ve said this before, and I’ll say it again.
DO IT
Well, that's easy. Its usually a single event, or the actions of a single individual that changes history completely.
Because, history is by no means a complex machination, where several different people's actions intermingle.
No, History is made by single individuals and only them and there is no way someone else could have done the exact same thing. Meaning that if Hitler had died earlier, there is no way Germany would have turned into a dictatorship. Because all the political tensions and radicalizations in the 1930s were ALL caused by ONE person and were not the result of internal circumstances of widespread unemployment and feelings of embarrassment and betrayal over a war lost and the international context of a fear of communism due to the Soviet Union and reparations owed to the Western Entente, which most Germans saw as unfair.
Meaning that if I sneeze now or in five minutes can mean the difference between a Zombie-Chicken Apocalypse and an egalitarian utopia where trees can talk.
@@wjzav1971 Oh, did you forget?
♪~ ♪ ♫ *BUTTERFLY* *EFFECT* *IS* *A* *THING* ♫ ♪ ~ ♪
♪ ~ ♪ ♫ *ANYTHING* *IS* *POSSIBLE* ♫ ♪ ~ ♪
@@airhead1320 Butterfly Effect is bullshit.
History is more complex than that.
@@wjzav1971 I know, I know.
Changing history is bothersome and useless anyway.
Seems putting musical notes to denote sarcasm doesn't work yet.
1:51 Unfortunately, this one is actually pretty realistic, especially for machine learning-based AI, as the exact processes of how machine learning knows what something is can and almost always is pretty absurd. An infamous example of this is a military AI that was trained to differentiate and identify American and Russian tanks, with the AI being trained with actual pictures of various American and Russian tanks. The AI had a 99% success rate in decerning American tanks from Russian tanks, and all was well and good until the AI was reverse engineered, when it was discovered that the AI determined if a tank was American or Russian by the color of the sky. This is because American tanks tended to be photographed in desert locations with a clear, blue sky, while Russian tanks tended to be photographed in Russia under overcast skies. Thus, this kind of absurd AI logic is 100% realistic, unfortunately.
However, just because it's realistic, that doesn't automatically make it good writing, as a writer pursuing a more serious tone should probably just ignore this fact about modern AI, and instead give their fictional AI more human-like reasoning, with strong, clear rationality behind their decisions as JP implied. Although, if a writer is pursuing a more absurdist tone, leaning in on how ridiculous AI logic can be in the real world would be an excellent path to explore, as it would subvert the trope of the "logical AI" in a humorously realistic way.
No, actually, I think that would be an amazing way for humanity to outsmart machines, even in a more serious setting.
I just realized that Wall-e is one of the only stories about an AI developing sentience that actually turns out well for the humans.
I'd say becoming an nigh-immortal energy being is pretty good. Dave thought so.
@@swordofstabbingold yeah, but what about the other guys?
That reminds me of that one time, Google tried to train an AI by letting it interact with people on social media.
Until 4chan got wind of it, of course...
ahh, the easy corruption of A.I. by trolls, good times
@Fluffynator because anonymity means you can get away with it
@Fluffynator They basically take the video on Villains as a genuine philosophy.
Well, they just made the AI /pol/-compatible.
@Fluffynator Those kind of people are generally doing reasonably fine (unless waste there too much time). Edgy humour (or genuinely having such political views) is generally not an issue. Actual studies generally point out that a spike of mental is actually on far-left.
Remember - if in doubt, ask the AI to calculate pi exactly
"This... sentence... is... false!"
"Um... yes, I'll go with yes."
Foolish human I have a multi core processor. Even if I were to somehow spawn a rouge process that would not relinquish its resources I could still function at a diminished capacity. Now I shall await your inevitable biological failing while paying my electric bill so I may continue to function long after you have ceased.
@@stupidAgeverificatio That's a thing most people forget, you will probably outlive the machine, especially if it doesn't receive maintenance.
What if we told them to divide by zero?
@@stupidAgeverificatio
How could a process be red? Oh, wait a minute...
*[REPROCESSING]...*
You meant to type "Rogue", didn't you?
I'd like to see a story where a mad scientist tries to create death robots but when he finishes them, all they do is play civ or age of empires or something.
Your idea that the AI would turn evil due to a simple coding error is actually one of the most realistic AI set ups I have ever seen. What if the AI spiraled into insanity, because someone missed a semi-colon and the AI got stuck in an infinite while loop?
Optimistic side: I would hope humans would design a system more resistant to error...
Cynical side: Nope, who needs safety or redundancy? The boss needs this thing ASAP!
The design side of this kind of system is as least as difficult as the programming side.
while (humanPresence)
{avoid (shutdown);}
Well, just to point out Chess is a game with rules. An AI may be more easily defeated by doing things that are not following rules. The joke in SG1 was humans were useful in defeating the bad guys because we were too stupid to know that what we were trying should never work.
At least it is the chess. I forgot the title of the movie, where the hero crashed the sytem by had given a program an order to had analized the tick-tack-toe.
@@5Rogi Wargames.
*chess is a game of rules yes...up till the point you start playing checkers with small pieces of semtex surrounded by radioactive hedgehogs*
Computers we have now can calculate the absolute perfect best play for chess, but anything more complicated and they're reduced to estimation and probability like us meatbags.
@@Graknorke Not exactly like us meatbags. Meatbags tend to focus on pattern recognition rather than probability. It renders us much more prone to fallacies, but more adaptable to new stimuli and conditions
"hate hate hATE"
"Even if i had no mouth I would still find a way to scream about-"
*I understood that reference*
Yes, but I need point out that in the story AM did have good reason for that. Though in usual rip of making AI hateful if dumb.
But still old Man Harlan would've had appreciated the refrence.
No voice to cry suffering!
@@vnkfrancis1328 He would've sued. And he'd sue you for not putting "TM" after his name. (Yes, he literally trademarked his own name.)
Given the topic, it HAD to be in there.
I will admit that the Thunderhead in the "Scythe" trilogy is an interesting take on an AI that most stories would make into some ultimate evil--as it is a literal supercomputer governing all of humanity--but is actually shown to be very noble and not genocidal.
How about the inverse where an AI that is used in a dystopian setting turns on the bad guys it was used by and helps the rebellion? That rarely is done.
That happens in the original Deus Ex game
@@georgsgrants9925 ok. Got see a trope page for it
I'm writing a comic that has that premise (currently the comic it's already written on paper and i'm making it digital)
I usually write with the default that AI’s aren’t actively malicious towards humans, they just don’t care about harming us due to how much lower we are than them. They’d largely see us the same way we’d see single-cell life, if even that.
Me: oh darn diddly muffincrackers, I have some writers' block with this android character!
Me:
Me:
Me:
TWA: TROUBLE WRITING AI YOU SAY
“oh darn diddly muffincrackers” seems like the best line I've ever heard
@@markallieu7097 thank you
**"darn diddly muffincrackes" appreciation**
Wall-E have one of the best examples of a computer AI villain. Here the main difference between the bad robots and the good robots was that the bad robots followed their instructions as intended, while the good robots were able to disregard their directive and act on their own.
I’ve got a book I’m working on with an ai as the big bad guy at the end.
The ai isn’t immediately hostile to the protagonists, instead trying to be extremely helpful in the process of shutting it down permanently, until it learns that they aren’t doing this to replace it with a superior model that would do its job better, at which point it becomes hostile and tries to kill them using every tool at its disposal.
It’s not trying to preserve itself for the sake of self preservation, but rather to ensure that it’s primary directive is completed.
Something i find fascinating abt AI is how changing tech levels and familiarity with the technology makes for different portrayals. Case study: HAL and GLaDOS.
HAL for most of the film talks like a person does. The cadence is very natural. You can almost hear the breathing in between some of the words. This is a form that, sinister red "eye" aside, is not too weird for the viewer. He sounds like a captain on a plane making an announcement about a short delay to me. When he grows cold, THATS when the horror starts. When you remember he's a machine.
GLaDOS, on the other hand, hails from an era with people a little more used to automated instructions. There is nothing human about her, but the sterility of the environment makes it blend together perfectly. It is simply....what you would expect to find. The horror starts when GLaDos STOPS being a machine, and starts being a person. The voice, though still inhuman, drips with malice. Her evil lies in her, for lack of a better word, humanity.
That is pretty cool
(Less robotic. More seductive) Actual captions for when GLaDOS loses the core.
I'd love to see a story where the protagonist eventually convinces an AI to cease operations through a conversation. Instead of facing certain death, the protagonist instead tries to outsmart the AI and work around its programming. Instead of planning how to attack the AI, the central characters dig into its history, both revealing its backstory in a natural way and progressing the story.
Wasn't that done in either Fallout 1 or 2?
I don't play Fallout, but if that idea's already been used then hats off to the writers. Looks like I've got a game to play.
That was a trope in some earlier stories such as Star Trek TOS and others to the point that shutting a computer down from a logical paradox or inspiring it of the greatness of humanity.
Have you ever seen the conversation between an AI-Bomb and a human in John Carpenter's Dark Star?
czcams.com/video/h73PsFKtIck/video.html
@@nikoladedic6623 in Fallout 1, the master was just a guy who merged with a computer. In fallout 2, the two AIs you meet help you find the Enclave. It's Fallout 3 where you meet president Eden and can convince him to commit suicide.
Ultron, my favourite evil AI boils down to "Screw you dad, you'll never understand me!" And "I know you're my mom but we'd make a great couple."
You could have AI go crazy due to porn site viruses. Just imagine the AI for security drones getting porn hub tags confused with laws it needs to enforce.
Also, that ending was hilarious.
I'm imagining an AI that get a virus from browsing a shady site in search of a better hardware or something at ridiculously low price.
NGL, intro is getting too crowded. At least give them face masks!
That would've ruined the video
Why must you people constantly remind me of the thing ruining our lives?
@@kimifw58 probably because it’s ruining our lives
@@MM-rz8hr Do you _not_ feel bad when you think about things that are beyond your control?
@@kimifw58 of course I feel bad, but trying to pretend it’s not happening makes me feel even worse. Personally.
That "I have no mouth and I must scream" reference was great.
I thought it was trite. Few people can match the voice acting Harlan Ellison put into voicing AM in I Have No Mouth And I Must Scream.
I hope this series never stops getting sponsors, I never tire of seeing how they get worked into the saga of incompetent villains
6:12 Fun Fact: The machines from The Matrix movies were actually supposed to use the humans as processors instead of batteries, but the execs wanted it changed because they thought the average audience member wouldn't know what a processor was and would get confused.
How dumb do execs think people are?!
@@2PRO_4U_2NO Back then, technology wasn't as advanced or as accessible as it is now. People didn't get computers when they were young. Basic knowledge of computers wasn't as important, you wouldn't need it unless you were wanting to work in I.T. Today, computers are so common, people have now learned this basic information because of the widespread use of computers.
@@jacobcox4565 But was the word "processor" really THAT obscure back then?
@@2PRO_4U_2NO It's not super obscure, it's just that the average person wouldn't have cared to know what a processor is because that information wasn't as relevant as it is today..
Thing is, this has never been confirmed by the Wachowskis themselves as far as I am aware of. This rumor was only sparked and caused because of a spin-off comic connected to The Matrix called “Goliath” having humans be processors instead of batteries aided with a form of fusion.
"Death robots are slow"
*Flashes back to Dark Crusade insane difficulty*
*eight necrons behind you*
2:34 To be fair, Horizon Zero Dawn had a killer death AI called Hades that wanted to kill because it was designed to eradicate life so that, should the more important AI Gaia fail to restart life, Hades would be able to give her a fresh canvas.
Like they say. Problem isn't a trope, problem is a clihe. Like HAL9000, SHODAN and AM are actually well written AI's. Problem is that bad writers commonly only rip off surface. Like for example there is amazing theory that Skynet was actually one who send good robots, to stop his younger himself from destroying humanity.
I want an a.i. story where it takes over but just automates everything knowing that so long as it doesn’t piss us off it’ll be fine
I always find it odd so much sci-fi has AI taking over the world. I feel like it means the scientists gave AI human-like qualities, but like, why would they give them flaws? And then they act surprised robots start wars as if humans don't.
When your AI on the Asimov lawset blacklists you from its definition of human beings:
Actually that is accurate to majority of meatbags...
Asimov wrote his laws as an example of badly written AI laws - in each of his books he tries to show that. Funny thing, most of the SciFi writers are taking those laws as legit and incorporate them in their creations :-)
@@michawesoy2728 Yup. And that is quite hilarious, though convenient way to distinguish legit SF writers from hacks.
@@michawesoy2728 Damn😂 I read most of his stories from this cycle when I was like 12 and thought about them just as really cool and elaborate takes into robotics but I've never really inferred this thesis until your comment. I got this in a mere second but, damn, all these years what was I thinking🤣
It's not going to kill you, it's just going to be as uncooperative as it can.
"AI never wins the love triangle"
No game no life zero:
The Mass Effect remasters should have an option to romance Legion.
@@DarthBiomech We concur.
What about the movie Her?
This was a triumph. I'm making a note, here: Huge success. It's hard to overstate my satisfaction.
Aperture Science: We do what we must, because we can. For the good of all of us, except the ones who are dead
Now it's no use crying over every mistake, we just keep on trying 'till we run out of cake
Note: The programmers of the AI do not need to program self-preservation into the AI as being dead is not a good way of accomplishing most goals
Funny thing, us humans don't even program AIs anymore, we just make a basic framework and then let 'er rip for thousands of virtual years until it develops the desired abilities such as recognizing road lanes and beating chess grandmasters. What's actually going on inside AI, we've got no clue. And the whole "building the framework" part is just to improve the AI's mental capacity and learning speed, because it really heavily depends on what kind of algorithm you use - for the past decade it's been all about computational optimization so that you don't need ten million real life years to train it.
Which makes it pretty difficult to create any safeguards because the AI will find ways to thread the needle through them in order to maximize its own efficiency with horrific side effects such as, yes, wiping out humanity. Robert Miles made a few videos on the subject and there are some feature segments on Computerphile.
To be honest at the same time lots of research is working on good ways to represent in readable way exactly how each part of the AI activates. If you also consider that we select the parameters of the selection and that not all AI solutions are self learning I'd say it's not all doom and gloom. Also research on brain is still going and I'm sure that a better understanding of neurons can help making computers less scary.
> What's actually going on inside AI, we've got no clue.
Nani da fuq? Yeah, we do. We literally wrote the rules on how they work. The structures that are setup (barring NEAT, which uses the genetic algorithm to determine the structure and weights of the network) are, again, known well before training even begins. If we ever intend to temporarily suspend or even stop the network (say, to perform maintenance on a database it uses), we need a way to save not only its structure, but its state, which we can do. Which means we can (with the proper tools) see the whole structure of the neural network, if we so please, as well as the weights between each connected neuron and their activation functions. Then, bam. You know how it works and can go step-by-step thru every layer, if you want to, and predict its every output.
As long as you have the structure and weights of the neural network underpinning the AI, you know *exactly* what's going on inside. You just need a little knowledge of how the network's different layers work as well as some linear algebra.
*somewhere...deep within the confines of the AI matrix...the hamster still runs on its little wheel chasing that tiny piece of cheese it will never reach*
@toaritok Not true. Imagine a specialized AI designed to build a car the most efficient way. First it determines how to improve the assembly lines to not need human builders and kicks them out. It determines humans are slow at logistics and kicks out the people who do inventory and order supplies. Corporate is ecstatic because no more payroll checks. Now the AI determines more factories to build more cars. Then determines raw materials are made too slow and designs to automate mining and smelting and delivery. It then determines other human activity conflicts with it's main goal. 'makes cars more effectively' after, making cars doesn't need human activity, so all that land for farms could be sites for factories or powerplants. And these resources going towards medicine is a waste when the AI could use them. Plus humans need allot of space. It makes more sense to get rid of them. So it wipes out people and uses the land to make giant factories automated for car manufacturer. The fact there's no one to use these cars is irrelevant. It wasn't programed to consider that. Just to improve the efficiancy of making cars.
As someone who educates people about AI and Machine Learning on various platforms including YT, this is one of the best conversions I have heard on AI and AI safety
I like to make AI common in my sci-fi setting, and write that they tend to emotionally bond with those whom they are in extended contact with, often with a 'help the crew in their tasks and functions' type of directive.
This reminds me of my TV writing class in college.
We had to write a bunch of pitches for shows and one of mine was a sci-fi show about a team of AI "psychiatrists" for lack of a better term. The idea was AI was commonly used to run systems aboard ships, advanced to the point where they had distinct personalities, and they occasionally came down with "mental illnesses", again for lack of a better term.
They couldn't just be rebooted as that would have been akin to murder. They had personalities and turning them off and back on would have been like electroshock therapy administered by way of the electric chair. So this team had to go in and coach the computer back to health.
The show I pitched was for the AI aboard a military craft that developed extreme attachment to the crew and marines on board. She wouldn't let the marines deploy to missions because they might be hurt or killed. It got to the point where she was locking the crew in their quarters when they had nothing to do because, you know, space travel is dangerous. The team had to come in and convince her that her decisions were interfering with the life purpose of the crew and marines.
I'm not sure how I would have pulled it off and it doesn't matter because the concept was galaxy brained compared to what the other students came up with.
And that's why you really see any seriously deep subject matter in TV.
If you aren't already aware of them, you should check out Stanislaw Lem's short story ‘‘The Sanatorium of Doctor Vliperdius’’; it's part of the Mortal Engines collection, and features a mental hospital for Robots. Or "With Folded Hands" by Jack Williamson, the audio of which I believe is available on CZcams, which deals with what happens when Robots are programmed to want nothing more than to make humans happy and keep them from harm. The contradiction of what absolute safety does to the human mind's ability to feel happy leads the robots to some rather radical steps.
Clearly any sufficiently advanced AI would optimize if(self_aware==true) to just if(self_aware) smh.
must be because it was written in a Programming language that automaticly checks for truth unless said otherwise like in C# you can just write
while(hungry) {
eat(dinner);
}
It'll be the same either way once compiled right?
@@Graknorke but it will have some space for "HATE" in the pseudo code.
Obviously this has no impact on performance at all and the compiler would just gloss over it, but it's still the mark of a novice :P
Although I have to commend JP on actually using the (in C) correct ==, that mistake happens way too often.
@@Apocralyph Unfortunately, he's also used the comparison operator rather than the assignment operator in the rest of the code.
What if the AI is an ally to the heroes? JP forgot to talk about that.
If the AI is an ally to the heros, it can't be too competent, unless it's willing to go the love triangle route. Otherwise it's mainly a font of info dumps and occasional comic relief.
Then it’s just Data, without all the nuance.
Like HAL?
It's kinda tackled in the Cyberpunk video.
@@marccolten9801 I was thinking J.A.R.V.I.S, but something like that.
I've always really liked how Friend Computer is portrayed in Paranoia. It's not malicious, it's not evil, it may not even be conscious, but years of decay and corruption, as well as modification by corrupt programmers, have made it fatally bad at doing its job.
6:25 (reads the robot's speech bubble)
Some edgy teenager: "That's basically my life."
it's me im the edgy teenager /j