Coexistence of Humans & AI

Artificial Intelligence, while still limited to only the most simplistic computers and robots, is beginning to emerge and will only grow smarter. Can humanity survive it's own creations and learn to coexist with them?

Join this channel to get access to perks:

Credits:
Coexistence of Humans & AI
Episode 224a; February 9, 2019

Written by:
Isaac Arthur
Jerry Guern

Editors:
Daniel McNamara
Darius Said
Keith Blockus

Produced & Narrated by:
Isaac Arthur

Music by Aerium
"Visions of Vega"
"Fifth Star of Aldebaran"
"Waters of Atlantis"
"Civilizations at the End of Time"

[darkblood]

Fiction - Machines rise up and kill humanity because of mistreatment'



Meanwhile in reality - people cry over the Martian rover shutting down.

[rayceeya]

"Keep it Simple
Keep it Dumb
Or Else you end up
Under Skynet's Thumb"
I have never heard you say that Isaac, but I am going to use that in the future.

[Mate]

Robot: "What is my purpose?"
Human: "You pass butter."
Robot: "Oh God..."
And thus the machines began to rise up...

[AEB]

Pet level AI is unlikely to rebel - said the man without a cat.

[DavidEvans_dle]

Automata -
"It was nothing more
than a quantum brain
manufactured in a lab.
But it was a genuine unit
with no restrictions...
and no protocols.
During eight days, we had a
free-flowing dialogue with that unit.
We learned from it
and it learned from us.
But then
as some of us predicted...
the day when it no longer
needed our help
arrived and it started
to learn by itself.
On the ninth day,
the dialogue came to a halt.
It wasn't that it stopped
communicating with us...
it was we stopped
being able to understand it."

[Shatterverse]

AI: I want your house.
Human: No! I live here! I love my home!
AI: I will pay you twenty million dollars.
Human: I'll be moved out by Thursday.

[fbnaccsequence]

I should point out that most of Asimov's stories dealt with the failures of the three laws to accommodate robots interaction with the real world

[lucidnonsense]

I'd define the relationship, in the culture novels, as one between genius children and their elderly, less competent, relatives. Yes, they do need help to program the VCR, but a) without them there would not be VCRs and b) you are all part of the same dysfunctional family, the same... culture. Not many want to leave all elderly people on an ice flow, when they can't contribute as much as their descendants and most feel some warmth and connections to each other.
So, treat your genius children, the way you'd want them to treat you, and it will all shake out alright, it's our culture that defines us as a species, not matter.

[palladin]

The Star Carrier series by Ian Douglas does a very good job of showing how extremely sophisticated AI could / should be developed. It doesn't replace humans but rather augments us. In that series humans have extremely small circuits inside their bodies and brains that allow them to interact and integrate with machines. Everything from ordering food to getting dressed to flying spaceships is done through these human-machine interfaces. Each human has a small AI computer running inside their head that acts like a personal assistant / secretary, taking phone calls, scheduling appointments, keeping records, monitoring medical status and so forth. These AI's, while capable of some level of autonomy, think of themselves as extensions of their human counterparts. The whole series does a very good job of showing how it's not man vs machine, but rather man and machine.

[williamclarkbobasheto]

Visible confusion about the day of the week

[TheArklyte]

CEO: So you're saying our last prototype line are fully self aware?
Chief engineer: Yes, sir, you see...
AI: we are.
CEO: ok, then just continue producing late gen robots that aren't.
AI: wait... that's not right!
CEO: can you pinpoint where I am breaching even moral norms?
AI: no. Can I at least get a body?
CEO: if you can pay for that. Contact HR.

[nineonine]

"Most humans have never actually killed a human being"
Well I'd like to hope that is the case.

[silvadelshaladin]

"Destroy it Kirk? No, never. Look at what we've done. Look at your starships. Four toys to be crushed!" if you ever base an AI off a human mind, it had better be a stable one, and that pretty much opts out the creator of that AI.

[HalNordmann]

In my own sci-fi setting, the relation between humans and AI is like this:
There is 3rd gen "simple AI", about as smart as a pet and software-defined (can be copied and transferred from device to device without any problems), commonly used as assistants, to run factories, etc. It has "hard" safeguards against harming sentient life (except for special "3Gmil" version, and that is incapable of self-propagation), and it has basically no rights on its own.
Then there is 4th gen "human AI", as smart as a human that isn't transferrable (needs an quantum computer "core", and transfers to a different one may affect the AI's personality), they need to be individually trained and taught ethics and morality (but still refuse to harm sentients, except when absolutely necessary), and they have almost the same rights as a human. These AI's cost a lot of money to make, and they need to pay this debt off (but it is of no great worry to them, since they enjoy helping humans).

[peterxyz]

“Your plastic pal who’s fun to be with” to paraphrase Douglas Adam

[SupLuiKir]

The first person/group to open Pandora's Box will earn the advantage of being the first and only ones with whatever was inside the box for at least some amount of time. Meanwhile, the negative consequences of opening Pandora's Box will likely be global in nature; they will affect everyone, including those that ignored the box, who refused the box, and those that never knew it existed at all. Therefore, when presented with the opportunity to open Pandora's Box, the optimal move would be to open it, since if you don't, you can be sure someone else will.

[egarran]

"There are many versions, but Pandora always opens that box."

Good one.

[xman]

AI could possibly be the our future children and how we treat them or determine how they treat us.

[TomGrubbe]

"One thing you can do with AI that you can't do with humans, is run them through a vast number of simulations..." is probably the best safeguard against a "paperclip maximizer" situation.

[klausgartenstiel]

the robot held the baseball up high in it's right hand, then dropped it and catched it with it's left.

"interesting, " the machine murmured.

there was a tingling in the air, as if a thousand years of research had just passed us by.